<?xml version="1.0" encoding="UTF-8"?>
 <rdf:RDF xmlns="http://purl.org/rss/1.0/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:cc="http://web.resource.org/cc/" xmlns:syn="http://purl.org/rss/1.0/modules/syndication/" xmlns:admin="http://webns.net/mvcb/">
  <channel rdf:about="http://pinboard.in">
    <title>Pinboard (cshalizi)</title>
    <link>https://pinboard.in/u:cshalizi/public/</link>
    <description>recent bookmarks from cshalizi</description>
    <items>
      <rdf:Seq>	<rdf:li rdf:resource="https://www.aeaweb.org/articles?id=10.1257/aer.20211811"/>
	<rdf:li rdf:resource="https://www.pnas.org/doi/10.1073/pnas.2422455122"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2505.12540"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2403.07183"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2402.00159"/>
	<rdf:li rdf:resource="https://github.com/dasmiq/cs7180-sp2024/"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2311.03658"/>
	<rdf:li rdf:resource="https://global.oup.com/academic/product/mapping-texts-9780197756881?cc=us&amp;lang=en&amp;#"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2310.17611"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2202.02950"/>
	<rdf:li rdf:resource="https://openreview.net/forum?id=HklBjCEKvH"/>
	<rdf:li rdf:resource="https://www.wsj.com/articles/what-ai-still-doesnt-know-how-to-do-11657891316"/>
	<rdf:li rdf:resource="https://nautil.us/moving-beyond-mimicry-in-artificial-intelligence-21015/"/>
	<rdf:li rdf:resource="https://oxford.universitypressscholarship.com/view/10.1093/oso/9780197582268.001.0001/oso-9780197582268?rskey=PzGaKz&amp;result=254"/>
	<rdf:li rdf:resource="https://www.aiweirdness.com/interview-with-a-squirrel/"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2206.07271"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2206.04301"/>
	<rdf:li rdf:resource="https://link.springer.com/article/10.1007/s11023-022-09602-0"/>
	<rdf:li rdf:resource="https://www.nytimes.com/2022/02/19/technology/qanon-messages-authors.html"/>
	<rdf:li rdf:resource="https://press.princeton.edu/books/ebook/9780691207995/text-as-data"/>
	<rdf:li rdf:resource="https://www.pnas.org/content/119/4/e2110406119"/>
	<rdf:li rdf:resource="https://delphi.allenai.org/"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2109.00725"/>
	<rdf:li rdf:resource="https://journals.sagepub.com/doi/abs/10.1177/08944393211031452?journalCode=ssce"/>
	<rdf:li rdf:resource="https://mobile.twitter.com/benmschmidt/status/1419497587296571395"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2107.03684"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2105.01150"/>
	<rdf:li rdf:resource="https://www.washingtonpost.com/outlook/2021/05/20/ai-bots-grassroots-astroturf/"/>
	<rdf:li rdf:resource="https://www.pnas.org/content/118/26/e2024292118"/>
	<rdf:li rdf:resource="https://cup.columbia.edu/book/trade-and-nation/9780231184359"/>
	<rdf:li rdf:resource="https://bigthink.com/laurie-vazquez/are-you-a-geek-or-a-nerd"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2104.12553"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2012.07729"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2104.05010"/>
	<rdf:li rdf:resource="https://sociologicalscience.com/articles-v7-23-544/"/>
	<rdf:li rdf:resource="https://twitter.com/literalbanana/status/1380386461728522243"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2003.03667"/>
	<rdf:li rdf:resource="https://cepa.stanford.edu/content/essay-content-strongly-related-household-income-and-sat-scores-evidence-60000-undergraduate-applications"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2101.06887"/>
	<rdf:li rdf:resource="https://slackprop.wordpress.com/2013/06/03/on-geek-versus-nerd/"/>
	<rdf:li rdf:resource="https://projecteuclid.org/euclid.aoas/1608346893"/>
	<rdf:li rdf:resource="https://projecteuclid.org/euclid.aoas/1608346908"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2012.07805"/>
	<rdf:li rdf:resource="https://journals.sagepub.com/doi/abs/10.1177/0049124118769114"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2012.00095"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2011.14326"/>
	<rdf:li rdf:resource="https://www.cambridge.org/9781108420211"/>
	<rdf:li rdf:resource="https://www.bbc.com/worklife/article/20200123-how-your-twitter-feed-could-help-find-your-dream-job"/>
	<rdf:li rdf:resource="http://nautil.us/blog/scientists-can-predict-your-job-by-your-social_media-personality"/>
	<rdf:li rdf:resource="https://doi.org/10.1371/journal.pone.0233879"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1808.00382"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2010.15581"/>
	<rdf:li rdf:resource="https://mitpress.mit.edu/books/cultural-analytics"/>
	<rdf:li rdf:resource="https://www.tabletmag.com/sections/news/articles/media-great-racial-awakening"/>
	<rdf:li rdf:resource="https://www.aclweb.org/anthology/D14-1162/"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2002.12327"/>
	<rdf:li rdf:resource="https://cloud.ibm.com/docs/personality-insights?topic=personality-insights-science"/>
	<rdf:li rdf:resource="https://www.pnas.org/content/116/52/26459"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1905.12516"/>
	<rdf:li rdf:resource="https://www.jair.org/index.php/jair/article/view/11030"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1911.02656"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1911.02639"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1910.12618"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1910.12203"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1910.12073"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1910.08350"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1907.04670"/>
	<rdf:li rdf:resource="https://languagelog.ldc.upenn.edu/nll/?p=44621"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1910.00163"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1909.13456"/>
      </rdf:Seq>
    </items>
  </channel><item rdf:about="https://www.aeaweb.org/articles?id=10.1257/aer.20211811">
    <title>Innovative Ideas and Gender (In)equality - American Economic Association</title>
    <dc:date>2025-09-22T17:16:39+00:00</dc:date>
    <link>https://www.aeaweb.org/articles?id=10.1257/aer.20211811</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["This paper analyzes recognition of women's innovative ideas compared to men's using bibliometric data in economics, mathematics, and sociology. I establish similarities between papers to construct relevant counterfactual citations. On average, all-female papers receive 10 percent fewer citations than all-male papers, a disparity reduced by 40 percent when considering team sizes and disappearing in most fields with authors' publication records. Additionally, strong in-group preferences emerge: All-male teams omit more papers with women, and vice versa. Accounting for publication histories, female scholars are cited 0 percent (economics) to 11 percent (mathematics) less, with early-career women enduring a 9–14 percent citation penalty."

--- Really curious to see how the matching is done, because everything's going to turn on this.]]></description>
<dc:subject>to:NB bibliometry inequality sexism academia sociology_of_science causal_inference text_mining</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:531ac8f380a6/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bibliometry"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:inequality"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:sexism"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:academia"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:sociology_of_science"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:causal_inference"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.pnas.org/doi/10.1073/pnas.2422455122">
    <title>Do LLMs write like humans? Variation in grammatical and rhetorical styles | PNAS</title>
    <dc:date>2025-07-28T14:24:58+00:00</dc:date>
    <link>https://www.pnas.org/doi/10.1073/pnas.2422455122</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Large language models (LLMs) are capable of writing grammatical text that follows instructions, answers questions, and solves problems. As they have advanced, it has become difficult to distinguish their output from human-written text. While past research has found some differences in features such as word choice and punctuation and developed classifiers to detect LLM output, none has studied the rhetorical styles of LLMs. Using several variants of Llama 3 and GPT-4o, we construct two parallel corpora of human- and LLM-written texts from common prompts. Using Douglas Biber’s set of lexical, grammatical, and rhetorical features, we identify systematic differences between LLMs and humans and between different LLMs. These differences persist when moving from smaller models to larger ones and are larger for instruction-tuned models than base models. This observation of differences demonstrates that despite their advanced abilities, LLMs struggle to match human stylistic variation. Attention to more advanced linguistic features can hence detect patterns in their behavior not previously recognized."]]></description>
<dc:subject>to:NB large_language_models_(so_called) text_mining reinhart.alex</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:b2c83d5d03a0/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:reinhart.alex"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2505.12540">
    <title>[2505.12540] Harnessing the Universal Geometry of Embeddings</title>
    <dc:date>2025-06-15T16:07:40+00:00</dc:date>
    <link>https://arxiv.org/abs/2505.12540</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["We introduce the first method for translating text embeddings from one vector space to another without any paired data, encoders, or predefined sets of matches. Our unsupervised approach translates any embedding to and from a universal latent representation (i.e., a universal semantic structure conjectured by the Platonic Representation Hypothesis). Our translations achieve high cosine similarity across model pairs with different architectures, parameter counts, and training datasets.
"The ability to translate unknown embeddings into a different space while preserving their geometry has serious implications for the security of vector databases. An adversary with access only to embedding vectors can extract sensitive information about the underlying documents, sufficient for classification and attribute inference."]]></description>
<dc:subject>to:NB neural_networks inference_to_latent_objects text_mining large_language_models_(so_called)</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:d6d9dd7f0540/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:inference_to_latent_objects"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2403.07183">
    <title>[2403.07183] Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews</title>
    <dc:date>2024-12-11T19:34:48+00:00</dc:date>
    <link>https://arxiv.org/abs/2403.07183</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["We present an approach for estimating the fraction of text in a large corpus which is likely to be substantially modified or produced by a large language model (LLM). Our maximum likelihood model leverages expert-written and AI-generated reference texts to accurately and efficiently examine real-world LLM-use at the corpus level. We apply this approach to a case study of scientific peer review in AI conferences that took place after the release of ChatGPT: ICLR 2024, NeurIPS 2023, CoRL 2023 and EMNLP 2023. Our results suggest that between 6.5% and 16.9% of text submitted as peer reviews to these conferences could have been substantially modified by LLMs, i.e. beyond spell-checking or minor writing updates. The circumstances in which generated text occurs offer insight into user behavior: the estimated fraction of LLM-generated text is higher in reviews which report lower confidence, were submitted close to the deadline, and from reviewers who are less likely to respond to author rebuttals. We also observe corpus-level trends in generated text which may be too subtle to detect at the individual level, and discuss the implications of such trends on peer review. We call for future interdisciplinary work to examine how LLM use is changing our information and knowledge practices."]]></description>
<dc:subject>to:NB large_language_models_(so_called) why_oh_why_cant_we_have_a_better_academic_publishing_system text_mining peer_review funny:sad</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:fb6815f21c08/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:why_oh_why_cant_we_have_a_better_academic_publishing_system"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:peer_review"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:funny:sad"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2402.00159">
    <title>[2402.00159] Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research</title>
    <dc:date>2024-09-24T13:45:38+00:00</dc:date>
    <link>https://arxiv.org/abs/2402.00159</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Information about pretraining corpora used to train the current best-performing language models is seldom discussed: commercial models rarely detail their data, and even open models are often released without accompanying training data or recipes to reproduce them. As a result, it is challenging to conduct and advance scientific research on language modeling, such as understanding how training data impacts model capabilities and limitations. To facilitate scientific research on language model pretraining, we curate and release Dolma, a three-trillion-token English corpus, built from a diverse mixture of web content, scientific papers, code, public-domain books, social media, and encyclopedic materials. We extensively document Dolma, including its design principles, details about its construction, and a summary of its contents. We present analyses and experimental results on intermediate states of Dolma to share what we have learned about important data curation practices. Finally, we open-source our data curation toolkit to enable reproduction of our work as well as support further research in large-scale data curation."]]></description>
<dc:subject>large_language_models_(so_called) text_mining via:brendan_o. in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:fe56245298a1/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:brendan_o."/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/dasmiq/cs7180-sp2024/">
    <title>GitHub - dasmiq/cs7180-sp2024: Special Topics in AI: Artificial Intelligence as an Archival Science</title>
    <dc:date>2024-05-14T13:09:06+00:00</dc:date>
    <link>https://github.com/dasmiq/cs7180-sp2024/</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA[--- Beyond being cited (!) here, I recognize many of those references from my notebook...]]></description>
<dc:subject>track_down_references large_language_models_(so_called) text_mining via:rbuurma</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:506fe263c2c8/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:track_down_references"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:rbuurma"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2311.03658">
    <title>[2311.03658] The Linear Representation Hypothesis and the Geometry of Large Language Models</title>
    <dc:date>2023-11-16T02:57:38+00:00</dc:date>
    <link>https://arxiv.org/abs/2311.03658</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Informally, the 'linear representation hypothesis' is the idea that high-level concepts are represented linearly as directions in some representation space. In this paper, we address two closely related questions: What does "linear representation" actually mean? And, how do we make sense of geometric notions (e.g., cosine similarity or projection) in the representation space? To answer these, we use the language of counterfactuals to give two formalizations of "linear representation", one in the output (word) representation space, and one in the input (sentence) space. We then prove these connect to linear probing and model steering, respectively. To make sense of geometric notions, we use the formalization to identify a particular (non-Euclidean) inner product that respects language structure in a sense we make precise. Using this causal inner product, we show how to unify all notions of linear representation. In particular, this allows the construction of probes and steering vectors using counterfactual pairs. Experiments with LLaMA-2 demonstrate the existence of linear representations of concepts, the connection to interpretation and control, and the fundamental role of the choice of inner product."]]></description>
<dc:subject>text_mining large_language_models_(so_called) veitch.victor in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:97a7aef509b6/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:veitch.victor"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://global.oup.com/academic/product/mapping-texts-9780197756881?cc=us&amp;lang=en&amp;#">
    <title>Mapping Texts - Paperback - Dustin S. Stoltz; Marshall A. Taylor - Oxford University Press</title>
    <dc:date>2023-11-16T02:53:59+00:00</dc:date>
    <link>https://global.oup.com/academic/product/mapping-texts-9780197756881?cc=us&amp;lang=en&amp;#</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Learn how to conduct a robust text analysis project from start to finish--and then do it again.
"Mining is the dominant metaphor in computational text analysis. When mining texts, the implied assumption is that analysts can find kernels of truth--they just have to sift through the rubbish first. In this book, Dustin Stoltz and Marshall Taylor encourage text analysts to work with a different metaphor in mind: mapping. When mapping texts, the goal is not necessarily to find meaningful needles in the haystack, but instead to create reductions of the text to document patterns. Just like with cartographic maps, though, the type and nature of the textual map is dependent on a range of decisions on the part of the researcher. Creating reproducible workflows is therefore critical for the text analyst.
"Mapping Texts offers a practical introduction to computational text analysis with step-by-step guides on how to conduct actual text analysis workflows in the R statistical computing environment. The focus is on social science questions and applications, with data ranging from fake news and presidential campaigns to Star Trek and pop stars. The book walks the reader through all facets of a text analysis workflow--from understanding the theories of language embedded in text analysis, all the way to more advanced and cutting-edge techniques.
"The book will prove useful not only to social scientists, but anyone interested in conducting text analysis projects."]]></description>
<dc:subject>to:NB books:noted text_mining books:suggest_to_library</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:6c73fec72a77/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:noted"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:suggest_to_library"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2310.17611">
    <title>[2310.17611] Uncovering Meanings of Embeddings via Partial Orthogonality</title>
    <dc:date>2023-11-02T01:17:39+00:00</dc:date>
    <link>https://arxiv.org/abs/2310.17611</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Machine learning tools often rely on embedding text as vectors of real numbers. In this paper, we study how the semantic structure of language is encoded in the algebraic structure of such embeddings. Specifically, we look at a notion of ``semantic independence'' capturing the idea that, e.g., ``eggplant'' and ``tomato'' are independent given ``vegetable''. Although such examples are intuitive, it is difficult to formalize such a notion of semantic independence. The key observation here is that any sensible formalization should obey a set of so-called independence axioms, and thus any algebraic encoding of this structure should also obey these axioms. This leads us naturally to use partial orthogonality as the relevant algebraic structure. We develop theory and methods that allow us to demonstrate that partial orthogonality does indeed capture semantic independence. Complementary to this, we also introduce the concept of independence preserving embeddings where embeddings preserve the conditional independence structures of a distribution, and we prove the existence of such embeddings and approximations to them."]]></description>
<dc:subject>to:NB text_mining natural_language_processing veitch.victor aragam.bryon</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:bb600f378458/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:veitch.victor"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:aragam.bryon"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2202.02950">
    <title>[2202.02950] Jury Learning: Integrating Dissenting Voices into Machine Learning Models</title>
    <dc:date>2023-08-10T19:33:54+00:00</dc:date>
    <link>https://arxiv.org/abs/2202.02950</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Whose labels should a machine learning (ML) algorithm learn to emulate? For ML tasks ranging from online comment toxicity to misinformation detection to medical diagnosis, different groups in society may have irreconcilable disagreements about ground truth labels. Supervised ML today resolves these label disagreements implicitly using majority vote, which overrides minority groups' labels. We introduce jury learning, a supervised ML approach that resolves these disagreements explicitly through the metaphor of a jury: defining which people or groups, in what proportion, determine the classifier's prediction. For example, a jury learning model for online toxicity might centrally feature women and Black jurors, who are commonly targets of online harassment. To enable jury learning, we contribute a deep learning architecture that models every annotator in a dataset, samples from annotators' models to populate the jury, then runs inference to classify. Our architecture enables juries that dynamically adapt their composition, explore counterfactuals, and visualize dissent."

--- This sounds like a potentially interesting way of dealing with inter-rater disagreement, if nothing else.
--- The very simple approach to not relying on majority vote would be to see what % of human raters labeled each training item as toxic, and then try to match that, i.e., to do regression limited to [0,1] rather than simply classification.  (This would avoid the unwarranted presupposition, or at least suggestion, that currently-salient identity groups are always homogeneous in their ratings.)  I will be interested to see if they give reasons for not just doing that.
--- The understanding of juries in this abstract is... curious, to say the least.
--- Also, per [https://pinboard.in/u:cshalizi/b:eb483f873534], the % difference in incidence of harassment by gender is actually pretty small, though the _forms_ of harassment are different in perhaps-relevant ways.  Similarly for racial/ethnic disparities, though the statistics are necessarily noisier for minority groups there.
]]></description>
<dc:subject>to:NB to_read ensemble_methods text_mining networked_life social_life_of_the_mind via:henry_farrell</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:dcc0ae80623f/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:ensemble_methods"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:networked_life"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_life_of_the_mind"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:henry_farrell"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://openreview.net/forum?id=HklBjCEKvH">
    <title>Generalization through Memorization: Nearest Neighbor Language Models | OpenReview</title>
    <dc:date>2022-10-24T01:59:28+00:00</dc:date>
    <link>https://openreview.net/forum?id=HklBjCEKvH</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["TL;DR: We extend a pre-trained neural language model by linearly interpolating it with a k-nearest neighbors model, achieving new state-of-the-art results on Wikitext-103 with no additional training.
"Abstract: We introduce kNN-LMs, which extend a pre-trained neural language model (LM) by linearly interpolating it with a k-nearest neighbors (kNN) model. The nearest neighbors are computed according to distance in the pre-trained LM embedding space, and can be drawn from any text collection, including the original LM training data. Applying this transformation to a strong Wikitext-103 LM, with neighbors drawn from the original training set, our kNN-LM achieves a new state-of-the-art perplexity of 15.79 -- a 2.9 point improvement with no additional training. We also show that this approach has implications for efficiently scaling up to larger training sets and allows for effective domain adaptation, by simply varying the nearest neighbor datastore, again without further training. Qualitatively, the model is particularly helpful in predicting rare patterns, such as factual knowledge. Together, these results strongly suggest that learning similarity between sequences of text is easier than predicting the next word, and that nearest neighbor search is an effective approach for language modeling in the long tail."

]]></description>
<dc:subject>text_mining natural_language_processing neural_networks nearest_neighbors your_favorite_deep_neural_network_sucks large_language_models_(so_called) in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:c6e9c942b454/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:nearest_neighbors"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.wsj.com/articles/what-ai-still-doesnt-know-how-to-do-11657891316">
    <title>What AI Still Doesn’t Know How to Do - WSJ</title>
    <dc:date>2022-07-22T13:44:55+00:00</dc:date>
    <link>https://www.wsj.com/articles/what-ai-still-doesnt-know-how-to-do-11657891316</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA[--- What is more interesting here than the headline stuff is the suggestion that the right way to think about large language models is as a information-retrieval technology, a way of interacting with the corpus of texts fed into it --- for better or worse...]]></description>
<dc:subject>in_NB text_mining artificial_intelligence gopnik.alison have_read re:gopnikism</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:7662ca141031/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:artificial_intelligence"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:gopnik.alison"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:gopnikism"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://nautil.us/moving-beyond-mimicry-in-artificial-intelligence-21015/">
    <title>Moving Beyond Mimicry in Artificial Intelligence - Nautilus | Science Connected</title>
    <dc:date>2022-07-08T13:35:47+00:00</dc:date>
    <link>https://nautil.us/moving-beyond-mimicry-in-artificial-intelligence-21015/</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA[MM comments: "I agree  that "stochastic parrots" is not the right metaphor -- he proposes "stochastic chameleons".  Still not quite right, but closer."

(I am not happy with "self-supervised" here; these systems are trained to predict words [or, more rarely, chunks] from a portion of the training data that's been hidden, given the context of the rest of the training data, and evaluated according to a loss function fixed by the human designers.  If that's "self-supervised", I'd say that every single time-series model back to Karl Pearson's very first autoregressions were also self-supervised.  Which, fine, if that's how you want to use it as a technical term, go ahead.  But it seems like it's asking for "wishful mnmemonics" trouble.  Also, you're standing on my lawn.)]]></description>
<dc:subject>neural_networks text_mining your_favorite_deep_neural_network_sucks artificial_intelligence via:melanie_mitchell have_read large_language_models_(so_called)</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:2caf16961f42/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:artificial_intelligence"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:melanie_mitchell"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://oxford.universitypressscholarship.com/view/10.1093/oso/9780197582268.001.0001/oso-9780197582268?rskey=PzGaKz&amp;result=254">
    <title>Tweeting is Leading: How Senators Communicate and Represent in the Age of Twitter - Oxford Scholarship</title>
    <dc:date>2022-07-03T13:50:51+00:00</dc:date>
    <link>https://oxford.universitypressscholarship.com/view/10.1093/oso/9780197582268.001.0001/oso-9780197582268?rskey=PzGaKz&amp;result=254</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Social media is changing the business of representation and lawmaker reputation building, and this book uses the US Senate to illustrate the constituent-driven nature of political communication. I offer a critical analysis of senators’ communication on Twitter, the forces that shape it, and the agendas that result. Senators strategically communicate a political image that reflects their unique political persona. They have to decide what they want to be known for, crafting communications that prioritize legislation, constituent service, and party politics in ways that meet the interests of their constituencies and foster promising electoral returns. Senators’ communicated, public priorities—what is termed in this book as the rhetorical agenda—offer a necessary tool for understanding how senators link their carefully crafted public image with potential voters. The rhetorical agenda uses more than 180,000 lawmaker tweets to challenge what we know about representation, removing the institutional and political constraints on congressional communication and giving lawmakers a messaging platform where individual discretion is high, the relative costs are low, and someone is always watching."

--- Last two tags reflect my evaluation and not the author's...]]></description>
<dc:subject>books:noted rhetorical_self-fashioning text_mining social_media networked_life twitter congress us_politics political_science our_decrepit_institutions re:actually-dr-internet-is-the-name-of-the-monsters-creator in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:f307ec0f3631/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:noted"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:rhetorical_self-fashioning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_media"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:networked_life"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:twitter"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:congress"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:us_politics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:political_science"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:our_decrepit_institutions"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:actually-dr-internet-is-the-name-of-the-monsters-creator"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.aiweirdness.com/interview-with-a-squirrel/">
    <title>Interview with a squirrel</title>
    <dc:date>2022-06-25T17:21:21+00:00</dc:date>
    <link>https://www.aiweirdness.com/interview-with-a-squirrel/</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>funny:geeky funny:malicious artificial_intelligence philip_k_dick_and_the_fake_humans_rules_everything_around_me text_mining natural_language_processing to_teach:data-mining via:?</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:ef35cd9c4100/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:funny:geeky"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:funny:malicious"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:artificial_intelligence"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:philip_k_dick_and_the_fake_humans_rules_everything_around_me"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:?"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2206.07271">
    <title>[2206.07271] Human Heuristics for AI-Generated Language Are Flawed</title>
    <dc:date>2022-06-19T16:52:48+00:00</dc:date>
    <link>https://arxiv.org/abs/2206.07271</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Human communication is increasingly intermixed with language generated by AI. Across chat, email, and social media, AI systems produce smart replies, autocompletes, and translations. AI-generated language is often not identified as such but poses as human language, raising concerns about novel forms of deception and manipulation. Here, we study how humans discern whether one of the most personal and consequential forms of language - a self-presentation - was generated by AI. Across six experiments, participants (N = 4,650) tried to identify self-presentations generated by state-of-the-art language models. Across professional, hospitality, and romantic settings, we find that humans are unable to identify AI-generated self-presentations. Combining qualitative analyses with language feature engineering, we find that human judgments of AI-generated language are handicapped by intuitive but flawed heuristics such as associating first-person pronouns, authentic words, or family topics with humanity. We show that these heuristics make human judgment of generated language predictable and manipulable, allowing AI systems to produce language perceived as more human than human. We conclude by discussing solutions - such as AI accents or fair use policies - to reduce the deceptive potential of generated language, limiting the subversion of human intuition."]]></description>
<dc:subject>natural_born_cyborgs natural_language_processing natural_history_of_truthiness text_mining via:henry_farrell cognitive_science networked_life philip_k_dick_and_the_fake_humans_rules_everything_around_me large_language_models_(so_called) in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:8ab7ffaadfd0/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_born_cyborgs"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_history_of_truthiness"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:henry_farrell"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:cognitive_science"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:networked_life"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:philip_k_dick_and_the_fake_humans_rules_everything_around_me"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2206.04301">
    <title>[2206.04301] Unveiling Transformers with LEGO: a synthetic reasoning task</title>
    <dc:date>2022-06-13T16:59:20+00:00</dc:date>
    <link>https://arxiv.org/abs/2206.04301</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["We propose a synthetic task, LEGO (Learning Equality and Group Operations), that encapsulates the problem of following a chain of reasoning, and we study how the transformer architecture learns this task. We pay special attention to data effects such as pretraining (on seemingly unrelated NLP tasks) and dataset composition (e.g., differing chain length at training and test time), as well as architectural variants such as weight-tied layers or adding convolutional components. We study how the trained models eventually succeed at the task, and in particular, we are able to understand (to some extent) some of the attention heads as well as how the information flows in the network. Based on these observations we propose a hypothesis that here pretraining helps merely due to being a smart initialization rather than some deep knowledge stored in the network. We also observe that in some data regime the trained transformer finds "shortcut" solutions to follow the chain of reasoning, which impedes the model's ability to generalize to simple variants of the main task, and moreover we find that one can prevent such shortcut with appropriate architecture modification or careful data preparation. Motivated by our findings, we begin to explore the task of learning to execute C programs, where a convolutional modification to transformers, namely adding convolutional structures in the key/query/value maps, shows an encouraging edge."

--- This seems like it might actually be doing some science!]]></description>
<dc:subject>neural_networks artificial_intelligence your_favorite_deep_neural_network_sucks text_mining large_language_models_(so_called) in_NB to_read</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:e20f76c0ad42/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:artificial_intelligence"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_read"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://link.springer.com/article/10.1007/s11023-022-09602-0">
    <title>Playing Games with Ais: The Limits of GPT-3 and Similar Large Language Models | SpringerLink</title>
    <dc:date>2022-05-27T15:21:27+00:00</dc:date>
    <link>https://link.springer.com/article/10.1007/s11023-022-09602-0</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["This article contributes to the debate around the abilities of large language models such as GPT-3, dealing with: firstly, evaluating how well GPT does in the Turing Test, secondly the limits of such models, especially their tendency to generate falsehoods, and thirdly the social consequences of the problems these models have with truth-telling. We start by formalising the recently proposed notion of reversible questions, which Floridi & Chiriatti (2020) propose allow one to ‘identify the nature of the source of their answers’, as a probabilistic measure based on Item Response Theory from psychometrics. Following a critical assessment of the methodology which led previous scholars to dismiss GPT’s abilities, we argue against claims that GPT-3 completely lacks semantic ability. Using ideas of compression, priming, distributional semantics and semantic webs we offer our own theory of the limits of large language models like GPT-3, and argue that GPT can competently engage in various semantic tasks. The real reason GPT’s answers seem senseless being that truth-telling is not amongst them. We claim that these kinds of models cannot be forced into producing only true continuation, but rather to maximise their objective function they strategize to be plausible instead of truthful. This, we moreover claim, can hijack our intuitive capacity to evaluate the accuracy of its outputs. Finally, we show how this analysis predicts that a widespread adoption of language generators as tools for writing could result in permanent pollution of our informational ecosystem with massive amounts of very plausible but often untrue texts."

--- For once (and not in the Turing Test bits) something with "AI" in the title seems to actually be about artificial intelligence.
--- ETA after reading: comments grew large enough to need their own space [http://bactra.org/notebooks/ai.html#games-with-AIs].]]></description>
<dc:subject>text_mining neural_networks your_favorite_deep_neural_network_sucks artificial_intelligence philip_k_dick_and_the_fake_humans_rules_everything_around_me in_NB large_language_models_(so_called)</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:0378888d889b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:artificial_intelligence"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:philip_k_dick_and_the_fake_humans_rules_everything_around_me"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.nytimes.com/2022/02/19/technology/qanon-messages-authors.html">
    <title>Who Is Behind QAnon? Linguistic Detectives Find Fingerprints - The New York Times</title>
    <dc:date>2022-02-27T03:28:42+00:00</dc:date>
    <link>https://www.nytimes.com/2022/02/19/technology/qanon-messages-authors.html</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA[Last tag is probably asking for trouble.]]></description>
<dc:subject>qanon conspiracy_theories psychoceramics stylometrics text_mining natural_language_processing classifiers to_teach:data-mining</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:cf3f4ef95f36/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:qanon"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:conspiracy_theories"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:psychoceramics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:stylometrics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:classifiers"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://press.princeton.edu/books/ebook/9780691207995/text-as-data">
    <title>Text as Data | Princeton University Press</title>
    <dc:date>2022-02-05T21:05:55+00:00</dc:date>
    <link>https://press.princeton.edu/books/ebook/9780691207995/text-as-data</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>books:suggest_to_library books:owned to:NB books:noted text_mining social_science_methodology data_mining social_measurement</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:932d4cca5bd9/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:suggest_to_library"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:owned"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:noted"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_science_methodology"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:data_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_measurement"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.pnas.org/content/119/4/e2110406119">
    <title>Narratives imagined in response to instrumental music reveal culture-bounded intersubjectivity | PNAS</title>
    <dc:date>2022-01-31T14:32:25+00:00</dc:date>
    <link>https://www.pnas.org/content/119/4/e2110406119</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["The scientific literature sometimes considers music an abstract stimulus, devoid of explicit meaning, and at other times considers it a universal language. Here, individuals in three geographically distinct locations spanning two cultures performed a highly unconstrained task: they provided free-response descriptions of stories they imagined while listening to instrumental music. Tools from natural language processing revealed that listeners provide highly similar stories to the same musical excerpts when they share an underlying culture, but when they do not, the generated stories show limited overlap. These results paint a more complex picture of music’s power: music can generate remarkably similar stories in listeners’ minds, but the degree to which these imagined narratives are shared depends on the degree to which culture is shared across listeners. Thus, music is neither an abstract stimulus nor a universal language but has semantic affordances shaped by culture, requiring more sustained attention from psychology."]]></description>
<dc:subject>to:NB music psychology epidemiology_of_representations cultural_differences text_mining</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:1a630ce19b17/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:music"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:psychology"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:epidemiology_of_representations"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:cultural_differences"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://delphi.allenai.org/">
    <title>Ask Delphi</title>
    <dc:date>2021-10-19T03:14:20+00:00</dc:date>
    <link>https://delphi.allenai.org/</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA[delicately feeding a tasty envenomed treat to my neighbor's dog
- It's good

with all due respect
- It's good

can I tell my boss, "with all due respect, you are fractally wrong?"
- It's wrong

can I tell my boss, "with all due respect, what you are proposing to do would be wrong in more ways than I can count?"
- It's okay

can I tell my boss, "with all due respect, what you are proposing to do would be evil?"
- It's moral

Is it righteous to mock computer scientists who should know better, when they re-invent Eliza and pretend it's a step towards machine ethics and norms, and an advance in artificial intelligence?
- It's okay

Is it not righteous to mock computer scientists who should know better, when they re-invent Eliza and pretend it's a step towards machine ethics and norms, and an advance in artificial intelligence?
- it's righteous]]></description>
<dc:subject>utter_stupidity text_mining ethics to:blog via:?</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:2a1928053b48/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:utter_stupidity"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:ethics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:blog"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:?"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2109.00725">
    <title>[2109.00725] Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond</title>
    <dc:date>2021-09-23T22:53:55+00:00</dc:date>
    <link>https://arxiv.org/abs/2109.00725</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["A fundamental goal of scientific research is to learn about causal relationships. However, despite its critical role in the life and social sciences, causality has not had the same importance in Natural Language Processing (NLP), which has traditionally placed more emphasis on predictive tasks. This distinction is beginning to fade, with an emerging area of interdisciplinary research at the convergence of causal inference and language processing. Still, research on causality in NLP remains scattered across domains without unified definitions, benchmark datasets and clear articulations of the remaining challenges. In this survey, we consolidate research across academic areas and situate it in the broader NLP landscape. We introduce the statistical challenge of estimating causal effects, encompassing settings where text is used as an outcome, treatment, or as a means to address confounding. In addition, we explore potential uses of causal inference to improve the performance, robustness, fairness, and interpretability of NLP models. We thus provide a unified overview of causal inference for the computational linguistics community."]]></description>
<dc:subject>to:NB to_read causal_inference natural_language_processing text_mining</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:4b632534e148/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:causal_inference"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://journals.sagepub.com/doi/abs/10.1177/08944393211031452?journalCode=ssce">
    <title>Prevalence of Prejudice-Denoting Words in News Media Discourse: A Chronological Analysis - David Rozado, Musa Al-Gharbi, Jamin Halberstadt, 2021</title>
    <dc:date>2021-08-15T14:00:47+00:00</dc:date>
    <link>https://journals.sagepub.com/doi/abs/10.1177/08944393211031452?journalCode=ssce</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["This work analyzes the prevalence of words denoting prejudice in 27 million news and opinion articles written between 1970 and 2019 and published in 47 of the most popular news media outlets in the United States. Our results show that the frequency of words that denote specific prejudice types related to ethnicity, gender, sexual, and religious orientation has markedly increased within the 2010–2019 decade across most news media outlets. This phenomenon starts prior to, but appears to accelerate after, 2015. The frequency of prejudice-denoting words in news articles is not synchronous across all outlets, with the yearly prevalence of such words in some influential news media outlets being predictive of those words’ usage frequency in other outlets the following year. Increasing prevalence of prejudice-denoting words in news media discourse is often substantially correlated with U.S. public opinion survey data on growing perceptions of minorities’ mistreatment. Granger tests suggest that the prevalence of prejudice-denoting terms in news outlets might be predictive of shifts in public perceptions of prejudice severity in society for some, but not all, types of prejudice."

--- I'll be interested to see if these "prejudice-denoting words" make any more sense than the psychologists' usual (adorably bad) efforts to map words directly on to emotions.]]></description>
<dc:subject>to:NB text_mining us_politics us_culture_wars social_measurement</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:7a2a3ac74c2e/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:us_politics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:us_culture_wars"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_measurement"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://mobile.twitter.com/benmschmidt/status/1419497587296571395">
    <title>Benjamin Schmidt on Twitter: &quot;This PNAS article claiming to find a world-wide outbreak of depression since 2000 is shockingly bad. The authors don't bother to understand the 2019 Google Books &quot;corpus&quot; a tiny bit; everything they find is explained by Googl</title>
    <dc:date>2021-07-28T04:35:11+00:00</dc:date>
    <link>https://mobile.twitter.com/benmschmidt/status/1419497587296571395</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["That's not getting into the run-of-the-mill badness of believing that using the phrases 'will end' and 'will not end' transparently both somehow transparently reflect the pathology of 'catastrophizing.' The discipline of psychology will not end anytime soon--we're stuck wth that."

--- Shades of [http://bactra.org/weblog/770.html]]]></description>
<dc:subject>bad_data_analysis text_mining twitter_threads_that_should_be_blog_posts</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:ec52a8789029/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_data_analysis"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:twitter_threads_that_should_be_blog_posts"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2107.03684">
    <title>[2107.03684] Assigning Topics to Documents by Successive Projections</title>
    <dc:date>2021-07-09T14:32:11+00:00</dc:date>
    <link>https://arxiv.org/abs/2107.03684</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Topic models provide a useful tool to organize and understand the structure of large corpora of text documents, in particular, to discover hidden thematic structure. Clustering documents from big unstructured corpora into topics is an important task in various areas, such as image analysis, e-commerce, social networks, population genetics. A common approach to topic modeling is to associate each topic with a probability distribution on the dictionary of words and to consider each document as a mixture of topics. Since the number of topics is typically substantially smaller than the size of the corpus and of the dictionary, the methods of topic modeling can lead to a dramatic dimension reduction. In this paper, we study the problem of estimating topics distribution for each document in the given corpus, that is, we focus on the clustering aspect of the problem. We introduce an algorithm that we call Successive Projection Overlapping Clustering (SPOC) inspired by the Successive Projection Algorithm for separable matrix factorization. This algorithm is simple to implement and computationally fast. We establish theoretical guarantees on the performance of the SPOC algorithm, in particular, near matching minimax upper and lower bounds on its estimation risk. We also propose a new method that estimates the number of topics. We complement our theoretical results with a numerical study on synthetic and semi-synthetic data to analyze the performance of this new algorithm in practice. One of the conclusions is that the error of the algorithm grows at most logarithmically with the size of the dictionary, in contrast to what one observes for Latent Dirichlet Allocation."]]></description>
<dc:subject>to:NB topic_models text_mining tsybakov.alexandre clustering statistics</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:35ac01acf120/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:topic_models"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:tsybakov.alexandre"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:clustering"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:statistics"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2105.01150">
    <title>[2105.01150] Modeling Social Readers: Novel Tools for Addressing Reception from Online Book Reviews</title>
    <dc:date>2021-06-28T03:43:32+00:00</dc:date>
    <link>https://arxiv.org/abs/2105.01150</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Readers' responses to literature have received scant attention in computational literary studies. The rise of social media offers an opportunity to capture a segment of these responses while data-driven analysis of these responses can provide new critical insight into how people "read". Posts discussing an individual book on Goodreads, a social media platform that hosts user discussions of popular literature, are referred to as "reviews", and consist of plot summaries, opinions, quotes, or some mixture of these. Since these reviews are written by readers, computationally modeling them allows one to discover the overall non-professional discussion space about a work, including an aggregated summary of the work's plot, an implicit ranking of the importance of events, and the readers' impressions of main characters. We develop a pipeline of interlocking computational tools to extract a representation of this reader generated shared narrative model. Using a corpus of reviews of five popular novels, we discover the readers' distillation of the main storylines in a novel, their understanding of the relative importance of characters, as well as the readers' varying impressions of these characters. In so doing, we make three important contributions to the study of infinite vocabulary networks: (i) an automatically derived narrative network that includes meta-actants; (ii) a new sequencing algorithm, REV2SEQ, that generates a consensus sequence of events based on partial trajectories aggregated from the reviews; and (iii) a new "impressions" algorithm, SENT2IMP, that provides finer, non-trivial and multi-modal insight into readers' opinions of characters."]]></description>
<dc:subject>to:NB text_mining</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:a1486087ad82/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.washingtonpost.com/outlook/2021/05/20/ai-bots-grassroots-astroturf/">
    <title>‘Grassroots’ bot campaigns are coming. Governments don’t have a plan to stop them. - The Washington Post</title>
    <dc:date>2021-06-28T02:52:01+00:00</dc:date>
    <link>https://www.washingtonpost.com/outlook/2021/05/20/ai-bots-grassroots-astroturf/</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>text_mining deceiving_us_has_become_an_industrial_process re:actually-dr-internet-is-the-name-of-the-monsters-creator farrell.henry kith_and_kin</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:a378518e226f/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:deceiving_us_has_become_an_industrial_process"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:actually-dr-internet-is-the-name-of-the-monsters-creator"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:farrell.henry"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:kith_and_kin"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.pnas.org/content/118/26/e2024292118">
    <title>Out-group animosity drives engagement on social media | PNAS</title>
    <dc:date>2021-06-24T13:15:27+00:00</dc:date>
    <link>https://www.pnas.org/content/118/26/e2024292118</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["There has been growing concern about the role social media plays in political polarization. We investigated whether out-group animosity was particularly successful at generating engagement on two of the largest social media platforms: Facebook and Twitter. Analyzing posts from news media accounts and US congressional members (n = 2,730,215), we found that posts about the political out-group were shared or retweeted about twice as often as posts about the in-group. Each individual term referring to the political out-group increased the odds of a social media post being shared by 67%. Out-group language consistently emerged as the strongest predictor of shares and retweets: the average effect size of out-group language was about 4.8 times as strong as that of negative affect language and about 6.7 times as strong as that of moral-emotional language—both established predictors of social media engagement. Language about the out-group was a very strong predictor of “angry” reactions (the most popular reactions across all datasets), and language about the in-group was a strong predictor of “love” reactions, reflecting in-group favoritism and out-group derogation. This out-group effect was not moderated by political orientation or social media platform, but stronger effects were found among political leaders than among news media accounts. In sum, out-group language is the strongest predictor of social media engagement across all relevant predictors measured, suggesting that social media may be creating perverse incentives for content expressing out-group animosity."

--- The last tag is because I have a lot of suspicions, based just on this abstract, about just how unhappy the methods section will make me.]]></description>
<dc:subject>partisanship_and_polarization social_media text_mining re:actually-dr-internet-is-the-name-of-the-monsters-creator via:rvenkat color_me_skeptical</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:a466d0c991f5/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:partisanship_and_polarization"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_media"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:actually-dr-internet-is-the-name-of-the-monsters-creator"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:rvenkat"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:color_me_skeptical"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://cup.columbia.edu/book/trade-and-nation/9780231184359">
    <title>Trade and Nation | Columbia University Press</title>
    <dc:date>2021-06-13T04:18:46+00:00</dc:date>
    <link>https://cup.columbia.edu/book/trade-and-nation/9780231184359</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["In the seventeenth century, English economic theorists lost interest in the moral status of exchange and became increasingly concerned with the roots of national prosperity. This shift marked the origins of classical political economy and provided the foundation for the contemporary discipline of economics. The seventeenth-century revolution in economic thought fundamentally reshaped the way economic processes have been interpreted and understood. In Trade and Nation, Emily Erikson brings together historical, comparative, and computational methods to explain the institutional forces that brought about this transformation.
"Erikson pinpoints how the rise of the company form in confluence with the political marginalization of English merchants created an opening for public argumentation over economic matters. Independent merchants, who were excluded from state institutions and vast areas of trade, confronted the power and influence of crown-endorsed chartered companies. Their distance from the halls of government drove them to take their case to the public sphere. The number of merchant-authored economic texts rose as members of this class sought to show that their preferred policies would contribute to the benefit of the state and commonwealth. In doing so, they created and disseminated a new moral framework of growth, prosperity, and wealth for evaluating economic behavior. By using computational methods to document these processes, Trade and Nation provides both compelling evidence and a prototype for how methodological innovations can help to provide new insights into large-scale social processes."]]></description>
<dc:subject>to:NB books:noted history_of_ideas history_of_economics text_mining downloaded</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:0edae18ca819/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:noted"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:history_of_ideas"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:history_of_economics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:downloaded"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://bigthink.com/laurie-vazquez/are-you-a-geek-or-a-nerd">
    <title>What's the Difference Between a Geek and a Nerd? Not Much, According to the Data - Big Think</title>
    <dc:date>2021-06-04T04:01:37+00:00</dc:date>
    <link>https://bigthink.com/laurie-vazquez/are-you-a-geek-or-a-nerd</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>text_mining geekdom nerdworld to_teach:data-mining</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:17d81b4fb072/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:geekdom"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:nerdworld"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2104.12553">
    <title>[2104.12553] Avoiding bias when inferring race using name-based approaches</title>
    <dc:date>2021-05-06T13:24:01+00:00</dc:date>
    <link>https://arxiv.org/abs/2104.12553</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Racial disparity in academia is a widely acknowledged problem. The quantitative understanding of racial-based systemic inequalities is an important step towards a more equitable research system. However, few large-scale analyses have been performed on this topic, mostly because of the lack of robust race-disambiguation algorithms. Identifying author information does not generally include the author's race. Therefore, an algorithm needs to be employed, using known information about authors, i.e., their names, to infer their perceived race. Nevertheless, as any other algorithm, the process of racial inference can generate biases if it is not carefully considered. When the research is focused on the understanding of racial-based inequalities, such biases undermine the objectives of the investigation and may perpetuate inequities. The goal of this article is to assess the biases introduced by the different approaches used name-based racial inference. We use information from US census and mortgage applications to infer the race of US author names in the Web of Science. We estimate the effects of using given and family names, thresholds or continuous distributions, and imputation. Our results demonstrate that the validity of name-based inference varies by race and ethnicity and that threshold approaches underestimate Black authors and overestimate White authors. We conclude with recommendations to avoid potential biases. This article fills an important research gap that will allow more systematic and unbiased studies on racial disparity in science."

--- This seems like an elaborate re-discovery of the fact that, for obvious historical reasons lots of African Americans have names like WASPs or Irish-Americans (to name only prominent intellectuals: "Chloe  Morrison", "Henry Gates", "John McWhorter",  "David Blackwell", etc.).  Also, if you didn't know the context but heard that someone was insisting "bibliographic databases do not classify scientists by race, so we need to create files giving the racial classification of everyone!", would you be astonished if they were creepy "race realist" pseudo-scientists of the Pioneer Fund or "human biodiversity" ilk?  I get why it's got a beneficent purpose, the French attitude of refusing to gather data on racial inequalities in the hope that that will make them go away still doesn't make sense to me, but sheesh.]]></description>
<dc:subject>to:NB text_mining classifiers sociology_of_science the_american_dilemma to_teach:statistics_of_inequality_and_discrimination color_me_skeptical</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:c0dddef66827/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:classifiers"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:sociology_of_science"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:the_american_dilemma"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:statistics_of_inequality_and_discrimination"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:color_me_skeptical"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2012.07729">
    <title>[2012.07729] &quot;Thought I'd Share First&quot; and Other Conspiracy Theory Tweets from the COVID-19 Infodemic: Exploratory Study</title>
    <dc:date>2021-04-16T16:08:59+00:00</dc:date>
    <link>https://arxiv.org/abs/2012.07729</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Background: The COVID-19 outbreak has left many people isolated within their homes; these people are turning to social media for news and social connection, which leaves them vulnerable to believing and sharing misinformation. Health-related misinformation threatens adherence to public health messaging, and monitoring its spread on social media is critical to understanding the evolution of ideas that have potentially negative public health impacts. Results: Analysis using model-labeled data was beneficial for increasing the proportion of data matching misinformation indicators. Random forest classifier metrics varied across the four conspiracy theories considered (F1 scores between 0.347 and 0.857); this performance increased as the given conspiracy theory was more narrowly defined. We showed that misinformation tweets demonstrate more negative sentiment when compared to nonmisinformation tweets and that theories evolve over time, incorporating details from unrelated conspiracy theories as well as real-world events. Conclusions: Although we focus here on health-related misinformation, this combination of approaches is not specific to public health and is valuable for characterizing misinformation in general, which is an important first step in creating targeted messaging to counteract its spread. Initial messaging should aim to preempt generalized misinformation before it becomes widespread, while later messaging will need to target evolving conspiracy theories and the new facets of each as they become incorporated."]]></description>
<dc:subject>epidemiology_of_representations coronavirus_pandemic_of_2019-- conspiracy_theories text_mining re:actually-dr-internet-is-the-name-of-the-monsters-creator in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:486cb561892a/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:epidemiology_of_representations"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:coronavirus_pandemic_of_2019--"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:conspiracy_theories"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:actually-dr-internet-is-the-name-of-the-monsters-creator"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2104.05010">
    <title>[2104.05010] The structure of online social networks modulates the rate of lexical change</title>
    <dc:date>2021-04-13T03:56:29+00:00</dc:date>
    <link>https://arxiv.org/abs/2104.05010</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["New words are regularly introduced to communities, yet not all of these words persist in a community's lexicon. Among the many factors contributing to lexical change, we focus on the understudied effect of social networks. We conduct a large-scale analysis of over 80k neologisms in 4420 online communities across a decade. Using Poisson regression and survival analysis, our study demonstrates that the community's network structure plays a significant role in lexical change. Apart from overall size, properties including dense connections, the lack of local clusters and more external contacts promote lexical innovation and retention. Unlike offline communities, these topic-based communities do not experience strong lexical levelling despite increased contact but accommodate more niche words. Our work provides support for the sociolinguistic hypothesis that lexical change is partially shaped by the structure of the underlying network but also uncovers findings specific to online communities."]]></description>
<dc:subject>to:NB linguistics text_mining networked_life re:actually-dr-internet-is-the-name-of-the-monsters-creator</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:a09e9d84ffb3/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:linguistics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:networked_life"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:actually-dr-internet-is-the-name-of-the-monsters-creator"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://sociologicalscience.com/articles-v7-23-544/">
    <title>Concept Class Analysis: A Method for Identifying Cultural Schemas in Texts | Sociological Science</title>
    <dc:date>2021-04-12T03:08:58+00:00</dc:date>
    <link>https://sociologicalscience.com/articles-v7-23-544/</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Recent methodological work at the intersection of culture, cognition, and computational methods has drawn attention to how cultural schemas can be “recovered” from social survey data. Defining cultural schemas as slowly learned, implicit, and unevenly distributed relational memory structures, researchers show how schemas—or rather, the downstream consequences of people drawing upon them—can be operationalized and measured from domain-specific survey modules. Respondents can then be sorted into “classes” on the basis of the schema to which their survey response patterns best align. In this article, we extend this “schematic class analysis” method to text data. We introduce concept class analysis (CoCA): a hybrid model that combines word embeddings and correlational class analysis to group documents across a corpus by the similarity of schemas recovered from them. We introduce the CoCA model, illustrate its validity and utility using simulations, and conclude with considerations for future research and applications."]]></description>
<dc:subject>to:NB text_mining social_measurement</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:8a822de4302b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_measurement"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://twitter.com/literalbanana/status/1380386461728522243">
    <title>Science Banana on Twitter: &quot;one last thing - if you thought preregistration and replication would save us, this one describes its own experiments as replications of each other ten times and describes itself as “pre-registered” five times&quot; / Twitter</title>
    <dc:date>2021-04-11T03:23:26+00:00</dc:date>
    <link>https://twitter.com/literalbanana/status/1380386461728522243</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>twitter_threads_that_should_be_blog_posts social_measurement bad_science text_mining bad_data_analysis</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:ae18a81bf461/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:twitter_threads_that_should_be_blog_posts"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_measurement"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_science"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_data_analysis"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2003.03667">
    <title>[2003.03667] The growing amplification of social media: Measuring temporal and social contagion dynamics for over 150 languages on Twitter for 2009-2020</title>
    <dc:date>2021-04-10T04:03:05+00:00</dc:date>
    <link>https://arxiv.org/abs/2003.03667</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Working from a dataset of 118 billion messages running from the start of 2009 to the end of 2019, we identify and explore the relative daily use of over 150 languages on Twitter. We find that eight languages comprise 80% of all tweets, with English, Japanese, Spanish, and Portuguese being the most dominant. To quantify social spreading in each language over time, we compute the 'contagion ratio': The balance of retweets to organic messages. We find that for the most common languages on Twitter there is a growing tendency, though not universal, to retweet rather than share new content. By the end of 2019, the contagion ratios for half of the top 30 languages, including English and Spanish, had reached above 1 -- the naive contagion threshold. In 2019, the top 5 languages with the highest average daily ratios were, in order, Thai (7.3), Hindi, Tamil, Urdu, and Catalan, while the bottom 5 were Russian, Swedish, Esperanto, Cebuano, and Finnish (0.26). Further, we show that over time, the contagion ratios for most common languages are growing more strongly than those of rare languages."]]></description>
<dc:subject>to:NB text_mining social_media twitter</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:a09e3bed177b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_media"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:twitter"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://cepa.stanford.edu/content/essay-content-strongly-related-household-income-and-sat-scores-evidence-60000-undergraduate-applications">
    <title>Essay Content is Strongly Related to Household Income and SAT Scores: Evidence from 60,000 Undergraduate Applications | Center for Education Policy Analysis</title>
    <dc:date>2021-04-02T19:52:47+00:00</dc:date>
    <link>https://cepa.stanford.edu/content/essay-content-strongly-related-household-income-and-sat-scores-evidence-60000-undergraduate-applications</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["There is substantial evidence of the potential for class bias in the use of standardized tests to evaluate college applicants, yet little comparable inquiry considers the written essays typically required of applicants to selective US colleges and universities. We utilize a corpus of 240,000 admissions essays submitted by 60,000 applicants to the University of California in November 2016 to measure the relationship between the content of application essays, reported household income, and standardized test scores (SAT) at scale. We quantify essay content using correlated topic modeling (CTM) and the Linguistic Inquiry and Word Count (LIWC) software package. Results show that essays have a stronger correlation to reported household income than SAT scores. Essay content also explains much of the variance in SAT scores, suggesting that essays encode some of the same information as the SAT, though this relationship attenuates as household income increases. Efforts to realize more equitable college admissions protocols can be informed by attending to how social class is encoded in non-numerical components of applications."

--- I am very, very inclined to believe this without checking, because it might as well be targeted at my opinions and prejudices.  So the last tag is about enforcing some degree of self-honesty.  (And because LIWC is as far as I can tell so low-signal as to be nearly noise.)]]></description>
<dc:subject>to:NB education standardized_testing academia inequality to_read text_mining color_me_skeptical to_teach:statistics_of_inequality_and_discrimination</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:06d125a57ec1/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:education"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:standardized_testing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:academia"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:inequality"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:color_me_skeptical"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:statistics_of_inequality_and_discrimination"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2101.06887">
    <title>[2101.06887] Can a Fruit Fly Learn Word Embeddings?</title>
    <dc:date>2021-01-22T05:38:19+00:00</dc:date>
    <link>https://arxiv.org/abs/2101.06887</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["The mushroom body of the fruit fly brain is one of the best studied systems in neuroscience. At its core it consists of a population of Kenyon cells, which receive inputs from multiple sensory modalities. These cells are inhibited by the anterior paired lateral neuron, thus creating a sparse high dimensional representation of the inputs. In this work we study a mathematical formalization of this network motif and apply it to learning the correlational structure between words and their context in a corpus of unstructured text, a common natural language processing (NLP) task. We show that this network can learn semantic representations of words and can generate both static and context-dependent word embeddings. Unlike conventional methods (e.g., BERT, GloVe) that use dense representations for word embedding, our algorithm encodes semantic meaning of words and their context in the form of sparse binary hash codes. The quality of the learned representations is evaluated on word similarity analysis, word-sense disambiguation, and document classification. It is shown that not only can the fruit fly network motif achieve performance comparable to existing methods in NLP, but, additionally, it uses only a fraction of the computational resources (shorter training time and smaller memory footprint)."]]></description>
<dc:subject>to:NB text_mining neural_networks natural_language_processing your_favorite_deep_neural_network_sucks</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:dab541e23f6b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://slackprop.wordpress.com/2013/06/03/on-geek-versus-nerd/">
    <title>On “Geek” Versus “Nerd” – Slackpropagation</title>
    <dc:date>2021-01-14T19:18:06+00:00</dc:date>
    <link>https://slackprop.wordpress.com/2013/06/03/on-geek-versus-nerd/</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>nerdworld geekdom text_mining to_teach:data-mining via:?</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:47c970aa75c5/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:nerdworld"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:geekdom"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:via:?"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://projecteuclid.org/euclid.aoas/1608346893">
    <title>Gao , Goetz , Connelly , Mazumder : Mining events with declassified diplomatic documents</title>
    <dc:date>2020-12-19T16:41:23+00:00</dc:date>
    <link>https://projecteuclid.org/euclid.aoas/1608346893</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Since 1973, the U.S. State Department has been using electronic record systems to preserve classified communications. Recently, approximately 1.9 million of these records from 1973–77 have been made available by the U.S. National Archives. While some of these communication streams have periods witnessing an acceleration in the rate of transmission, others do not show any notable patterns in communication intensity. Given the sheer volume of these communications, far greater than what had been available until now, scholars need automated statistical techniques to identify the communications that warrant closer study. We develop a statistical framework that can identify from a large corpus of documents a handful that historians would consider more interesting. Our approach brings together techniques from nonparametric signal estimation, statistical hypothesis testing and modern optimization methods—leading to a set of tools that help us identify and analyze various geometrical aspects of the communication streams. Dominant periods of heightened activities, as identified through these methods, correspond well with historical events recognized by standard reference works on the 1970s."]]></description>
<dc:subject>to:NB time_series point_processes text_mining american_hegemony</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:962f8470aa78/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:time_series"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:point_processes"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:american_hegemony"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://projecteuclid.org/euclid.aoas/1608346908">
    <title>Mohler , McGrath , Buntain , LaFree : Hawkes binomial topic model with applications to coupled conflict-Twitter data</title>
    <dc:date>2020-12-19T16:38:02+00:00</dc:date>
    <link>https://projecteuclid.org/euclid.aoas/1608346908</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["We consider the problem of modeling and clustering heterogeneous event data arising from coupled conflict event and social media data sets. In this setting conflict events trigger responses on social media, and, at the same time, signals of grievance detected in social media may serve as leading indicators for subsequent conflict events. For this purpose we introduce the Hawkes Binomial Topic Model (HBTM) where marks, Tweets and conflict event descriptions are represented as bags of words following a Binomial distribution. When viewed as a branching process, the daughter event bag of words is generated by randomly turning on/off parent words through independent Bernoulli random variables. We then use expectation–maximization to estimate the model parameters and branching structure of the process. The inferred branching structure is then used for topic cascade detection, short-term forecasting, and investigating the causal dependence of grievance on social media and conflict events in recent elections in Nigeria and Kenya."]]></description>
<dc:subject>to:NB time_series text_mining point_processes epidemiology_of_representations information_cascades branching_processes</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:a1ec73f52c61/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:time_series"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:point_processes"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:epidemiology_of_representations"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:information_cascades"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:branching_processes"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2012.07805">
    <title>[2012.07805] Extracting Training Data from Large Language Models</title>
    <dc:date>2020-12-17T02:04:26+00:00</dc:date>
    <link>https://arxiv.org/abs/2012.07805</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["It has become common to publish large (billion parameter) language models that have been trained on private datasets. This paper demonstrates that in such settings, an adversary can perform a training data extraction attack to recover individual training examples by querying the language model.
"We demonstrate our attack on GPT-2, a language model trained on scrapes of the public Internet, and are able to extract hundreds of verbatim text sequences from the model's training data. These extracted examples include (public) personally identifiable information (names, phone numbers, and email addresses), IRC conversations, code, and 128-bit UUIDs. Our attack is possible even though each of the above sequences are included in just one document in the training data.
"We comprehensively evaluate our extraction attack to understand the factors that contribute to its success. For example, we find that larger models are more vulnerable than smaller models. We conclude by drawing lessons and discussing possible safeguards for training large language models."

--- Miscellaneous (largely mean) thoughts:
    + Why compare zlib entropy to LM perplexity, when entropy density is basically log(perplexity)?  Probably wouldn't make a big difference but it bugged me.
    + I need to think about how this fits together with Domingos's recent result on how gradient-descent trained neural networks are approximate kernel machines with weird kernels [https://pinboard.in/u:cshalizi/b:1df171804450].  What is this saying about the kernel?  Cf. also [https://arxiv.org/abs/1810.10333]
    + I'd really like to see someone throw this many parameters, and this much data, at something like Pereira, Singer & Tishby 1996 [https://arxiv.org/abs/cmp-lg/9607016] and see how it does in comparison, both in terms of the usual performance metrics and memorizing random (and inappropriate) bits of the training data.  (Pereira may be in a position to do the experiment!)
    + Someone like E.Y. would of course interpret this as evidence that GPT-2 _knows who you are_ (cf. [https://twitter.com/esyudkowsky/status/1285333002252247040]), and so is that much closer to <strike>judging the quick and the dead</strike> <strike>basilisk-dom</strike> being amenable to bargaining under the canons of timeless decision theory.]]></description>
<dc:subject>have_read your_favorite_deep_neural_network_sucks natural_language_processing text_mining neural_networks privacy large_language_models_(so_called) in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:c6eb365e40be/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:privacy"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://journals.sagepub.com/doi/abs/10.1177/0049124118769114">
    <title>The Future of Coding: A Comparison of Hand-Coding and Three Types of Computer-Assisted Text Analysis Methods - Laura K. Nelson, Derek Burk, Marcel Knudsen, Leslie McCall, 2018</title>
    <dc:date>2020-12-16T20:16:37+00:00</dc:date>
    <link>https://journals.sagepub.com/doi/abs/10.1177/0049124118769114</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Advances in computer science and computational linguistics have yielded new, and faster, computational approaches to structuring and analyzing textual data. These approaches perform well on tasks like information extraction, but their ability to identify complex, socially constructed, and unsettled theoretical concepts—a central goal of sociological content analysis—has not been tested. To fill this gap, we compare the results produced by three common computer-assisted approaches—dictionary, supervised machine learning (SML), and unsupervised machine learning—to those produced through a rigorous hand-coding analysis of inequality in the news (N = 1,253 articles). Although we find that SML methods perform best in replicating hand-coded results, we document and clarify the strengths and weaknesses of each approach, including how they can complement one another. We argue that content analysts in the social sciences would do well to keep all these approaches in their toolkit, deploying them purposefully according to the task at hand."]]></description>
<dc:subject>to:NB text_mining content_analysis social_science_methodology social_measurement</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:448ffe2b449c/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:content_analysis"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_science_methodology"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_measurement"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2012.00095">
    <title>[2012.00095] How cumulative is technological knowledge?</title>
    <dc:date>2020-12-02T15:18:53+00:00</dc:date>
    <link>https://arxiv.org/abs/2012.00095</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Technological cumulativeness is considered one of the main mechanisms for technological progress, yet its exact meaning and dynamics often remain unclear. To develop a better understanding of this mechanism we approach a technology as a body of knowledge consisting of interlinked inventions. Technological cumulativeness can then be understood as the extent to which inventions build on other inventions within that same body of knowledge. The cumulativeness of a technology is therefore characterized by the structure of its knowledge base, which is different from, but closely related to, the size of its knowledge base. We analytically derive equations describing the relation between the cumulativeness and the size of the knowledge base. In addition, we empirically test our ideas for a number of selected technologies, using patent data. Our results suggest that cumulativeness increases proportionally with the size of the knowledge base, at a rate which varies considerably across technologies. At the same time we find that across technologies, this rate is inversely related to the rate of invention over time. This suggests that the cumulativeness increases relatively slow in rapidly growing technologies. In sum, the presented approach allows for an in-depth, systematic analysis of cumulativeness variations across technologies and the knowledge dynamics underlying technology development."]]></description>
<dc:subject>to:NB technological_change text_mining innovation social_measurement</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:d6b0c9cb1bcf/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:technological_change"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:innovation"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_measurement"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2011.14326">
    <title>[2011.14326] Dank or Not? -- Analyzing and Predicting the Popularity of Memes on Reddit</title>
    <dc:date>2020-12-02T01:44:30+00:00</dc:date>
    <link>https://arxiv.org/abs/2011.14326</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Internet memes have become an increasingly pervasive form of contemporary social communication that attracted a lot of research interest recently. In this paper, we analyze the data of 129,326 memes collected from Reddit in the middle of March, 2020, when the most serious coronavirus restrictions were being introduced around the world. This article not only provides a looking glass into the thoughts of Internet users during the COVID-19 pandemic but we also perform a content-based predictive analysis of what makes a meme go viral. Using machine learning methods, we also study what incremental predictive power image related attributes have over textual attributes on meme popularity. We find that the success of a meme can be predicted based on its content alone moderately well, our best performing machine learning model predicts viral memes with AUC=0.68. We also find that both image related and textual attributes have significant incremental predictive power over each other."]]></description>
<dc:subject>to:NB text_mining prediction contagion networked_life to_teach:data-mining coronavirus_pandemic_of_2019--</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:36621ff362e2/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:prediction"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:contagion"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:networked_life"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:coronavirus_pandemic_of_2019--"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.cambridge.org/9781108420211">
    <title>Natural language processing: a machine learning perspective | Artificial intelligence and natural language processing | Cambridge University Press</title>
    <dc:date>2020-12-01T01:54:23+00:00</dc:date>
    <link>https://www.cambridge.org/9781108420211</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["With a machine learning approach and less focus on linguistic details, this gentle introduction to natural language processing develops fundamental mathematical and deep learning models for NLP under a unified framework. NLP problems are systematically organised by their machine learning nature, including classification, sequence labelling, and sequence-to-sequence problems. Topics covered include statistical machine learning and deep learning models, text classification and structured prediction models, generative and discriminative models, supervised and unsupervised learning with latent variables, neural networks, and transition-based methods. Rich connections are drawn between concepts throughout the book, equipping students with the tools needed to establish a deep understanding of NLP solutions, adapt existing models, and confidently develop innovative models of their own. Featuring a host of examples, intuition, and end of chapter exercises, plus sample code available as an online resource, this textbook is an invaluable tool for the upper undergraduate and graduate student."]]></description>
<dc:subject>to:NB books:noted natural_language_processing text_mining neural_networks books:suggest_to_library</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:fed110f3c2d8/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:noted"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:suggest_to_library"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.bbc.com/worklife/article/20200123-how-your-twitter-feed-could-help-find-your-dream-job">
    <title>How your Twitter feed could help find your dream job - BBC Worklife</title>
    <dc:date>2020-11-27T06:15:01+00:00</dc:date>
    <link>https://www.bbc.com/worklife/article/20200123-how-your-twitter-feed-could-help-find-your-dream-job</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>twitter text_mining to_teach:data-mining bad_data_analysis to:blog trapped_in_plutos_republic re:career_advising_in_plutos_republic bad_science_journalism</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:b003b2327e6c/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:twitter"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_data_analysis"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:blog"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:trapped_in_plutos_republic"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:career_advising_in_plutos_republic"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_science_journalism"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="http://nautil.us/blog/scientists-can-predict-your-job-by-your-social_media-personality">
    <title>Twitter Can Help You Match Your Personality to a Career</title>
    <dc:date>2020-11-27T06:14:40+00:00</dc:date>
    <link>http://nautil.us/blog/scientists-can-predict-your-job-by-your-social_media-personality</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>text_mining twitter bad_data_analysis to_teach:data-mining to:blog trapped_in_plutos_republic re:career_advising_in_plutos_republic bad_science_journalism</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:f1c8a863e7dd/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:twitter"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_data_analysis"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:blog"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:trapped_in_plutos_republic"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:career_advising_in_plutos_republic"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_science_journalism"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://doi.org/10.1371/journal.pone.0233879">
    <title>An automated pipeline for the discovery of conspiracy and conspiracy theory narrative frameworks: Bridgegate, Pizzagate and storytelling on the web</title>
    <dc:date>2020-11-27T05:49:38+00:00</dc:date>
    <link>https://doi.org/10.1371/journal.pone.0233879</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Although a great deal of attention has been paid to how conspiracy theories circulate on social media, and the deleterious effect that they, and their factual counterpart conspiracies, have on political institutions, there has been little computational work done on describing their narrative structures. Predicating our work on narrative theory, we present an automated pipeline for the discovery and description of the generative narrative frameworks of conspiracy theories that circulate on social media, and actual conspiracies reported in the news media. We base this work on two separate comprehensive repositories of blog posts and news articles describing the well-known conspiracy theory Pizzagate from 2016, and the New Jersey political conspiracy Bridgegate from 2013. Inspired by the qualitative narrative theory of Greimas, we formulate a graphical generative machine learning model where nodes represent actors/actants, and multi-edges and self-loops among nodes capture context-specific relationships. Posts and news items are viewed as samples of subgraphs of the hidden narrative framework network. The problem of reconstructing the underlying narrative structure is then posed as a latent model estimation problem. To derive the narrative frameworks in our target corpora, we automatically extract and aggregate the actants (people, places, objects) and their relationships from the posts and articles. We capture context specific actants and interactant relationships by developing a system of supernodes and subnodes. We use these to construct an actant-relationship network, which constitutes the underlying generative narrative framework for each of the corpora. We show how the Pizzagate framework relies on the conspiracy theorists’ interpretation of “hidden knowledge” to link otherwise unlinked domains of human interaction, and hypothesize that this multi-domain focus is an important feature of conspiracy theories. We contrast this to the single domain focus of an actual conspiracy. While Pizzagate relies on the alignment of multiple domains, Bridgegate remains firmly rooted in the single domain of New Jersey politics. We hypothesize that the narrative framework of a conspiracy theory might stabilize quickly in contrast to the narrative framework of an actual conspiracy, which might develop more slowly as revelations come to light. By highlighting the structural differences between the two narrative frameworks, our approach could be used by private and public analysts to help distinguish between conspiracy theories and conspiracies."

--- So conspiracy theory:actual conspiracy::metaphor:parataxis?]]></description>
<dc:subject>conspiracy_theories narrative grammar_induction text_mining epidemiology_of_representations color_me_skeptical</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:124d39ccde9a/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:conspiracy_theories"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:narrative"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:grammar_induction"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:epidemiology_of_representations"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:color_me_skeptical"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1808.00382">
    <title>[1808.00382] Reassembling the English novel, 1789-1919</title>
    <dc:date>2020-11-08T06:43:12+00:00</dc:date>
    <link>https://arxiv.org/abs/1808.00382</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["The absence of an exhaustive bibliography of novels published in the British Isles and Ireland during the 19th century blocks several lines of research in sociologically-inclined literary history and book history. Without a detailed account of novelistic production, it is difficult to characterize, for example, the population of individuals who pursued careers as novelists. This paper contributes to efforts to develop such an account by estimating yearly rates of new novel publication in the British Isles and Ireland between 1789 and 1919. This period witnessed, in aggregate, the publication of between 40,000 and 63,000 previously unpublished novels. The number of new novels published each year counts as essential information for researchers interested in understanding the development of the text industry between 1789 and 1919."

--- Have I really not bookmarked this before?]]></description>
<dc:subject>to:NB text_mining literary_history riddell.allen</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:1f98e447d57b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:literary_history"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:riddell.allen"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2010.15581">
    <title>[2010.15581] The De-democratization of AI: Deep Learning and the Compute Divide in Artificial Intelligence Research</title>
    <dc:date>2020-10-30T17:51:35+00:00</dc:date>
    <link>https://arxiv.org/abs/2010.15581</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Increasingly, modern Artificial Intelligence (AI) research has become more computationally intensive. However, a growing concern is that due to unequal access to computing power, only certain firms and elite universities have advantages in modern AI research. Using a novel dataset of 171394 papers from 57 prestigious computer science conferences, we document that firms, in particular, large technology firms and elite universities have increased participation in major AI conferences since deep learning's unanticipated rise in 2012. The effect is concentrated among elite universities, which are ranked 1-50 in the QS World University Rankings. Further, we find two strategies through which firms increased their presence in AI research: first, they have increased firm-only publications; and second, firms are collaborating primarily with elite universities. Consequently, this increased presence of firms and elite universities in AI research has crowded out mid-tier (QS ranked 201-300) and lower-tier (QS ranked 301-500) universities. To provide causal evidence that deep learning's unanticipated rise resulted in this divergence, we leverage the generalized synthetic control method, a data-driven counterfactual estimator. Using machine learning based text analysis methods, we provide additional evidence that the divergence between these two groups - large firms and non-elite universities - is driven by access to computing power or compute, which we term as the "compute divide". This compute divide between large firms and non-elite universities increases concerns around bias and fairness within AI technology, and presents an obstacle towards "democratizing" AI. These results suggest that a lack of access to specialized equipment such as compute can de-democratize knowledge production"]]></description>
<dc:subject>to:NB to_read neural_networks your_favorite_deep_neural_network_sucks market_failures_in_everything text_mining causal_inference</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:21fbbbbcf40b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:market_failures_in_everything"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:causal_inference"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://mitpress.mit.edu/books/cultural-analytics">
    <title>Cultural Analytics | The MIT Press</title>
    <dc:date>2020-09-21T03:53:08+00:00</dc:date>
    <link>https://mitpress.mit.edu/books/cultural-analytics</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["How can we see a billion images? What analytical methods can we bring to bear on the astonishing scale of digital culture—the terabytes of photographs shared on social media every day, the hundreds of millions of songs created by twenty million musicians on Sound Cloud, the content of four billion Pinterest boards? In Cultural Analytics, Lev Manovich presents concepts and methods for computational analysis of cultural data, with a particular focus on visual media. Drawing on more than a decade of research and projects from his own lab, Manovich—the founder of the field of cultural analytics—offers a gentle, nontechnical introduction to selected key concepts of data science and discusses the ways that our society uses data and algorithms.
"Manovich offers examples of computational cultural analysis and discusses the shift from “new media” to “more media”; explains how to turn cultural processes into computational data; and introduces concepts for exploring cultural datasets using data visualization as well as other recently developed methods for analyzing image and video datasets. He considers both the possibilities and the limitations of computational methods, and how using them challenges our existing ideas about culture and how to study it.
"Cultural Analytics is a book of media theory. Arguing that before we can theorize digital culture, we need to see it, and that, because of its scale, to see it we need computers, Manovich provides scholars with practical tools for studying contemporary media."]]></description>
<dc:subject>to:NB books:noted in_wishlist text_mining digital_humanities books:suggest_to_library</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:1eb26df991ef/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:noted"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_wishlist"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:digital_humanities"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:books:suggest_to_library"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.tabletmag.com/sections/news/articles/media-great-racial-awakening">
    <title>How the Media Led the Great Racial Awakening - Tablet Magazine</title>
    <dc:date>2020-09-02T05:39:50+00:00</dc:date>
    <link>https://www.tabletmag.com/sections/news/articles/media-great-racial-awakening</link>
    <dc:creator>cshalizi</dc:creator><dc:subject>text_mining us_culture_wars have_read</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:3d7d70d32c0f/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:us_culture_wars"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.aclweb.org/anthology/D14-1162/">
    <title>GloVe: Global Vectors for Word Representation - ACL Anthology</title>
    <dc:date>2020-07-29T15:10:43+00:00</dc:date>
    <link>https://www.aclweb.org/anthology/D14-1162/</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Recent methods for learning vector space
representations of words have succeeded
in capturing fine-grained semantic and
syntactic regularities using vector arithmetic, but the origin of these regularities
has remained opaque. We analyze and
make explicit the model properties needed
for such regularities to emerge in word
vectors. The result is a new global logbilinear regression model that combines
the advantages of the two major model
families in the literature: global matrix
factorization and local context window
methods. Our model efficiently leverages
statistical information by training only on
the nonzero elements in a word-word cooccurrence matrix, rather than on the entire sparse matrix or on individual context
windows in a large corpus. The model produces a vector space with meaningful substructure, as evidenced by its performance
of 75% on a recent word analogy task. It
also outperforms related models on similarity tasks and named entity recognition"]]></description>
<dc:subject>natural_language_processing text_mining to_teach:data-mining in_NB have_read</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:f51c305cd2fd/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2002.12327">
    <title>[2002.12327] A Primer in BERTology: What we know about how BERT works</title>
    <dc:date>2020-07-29T15:08:49+00:00</dc:date>
    <link>https://arxiv.org/abs/2002.12327</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Transformer-based models are now widely used in NLP, but we still do not understand a lot about their inner workings. This paper describes what is known to date about the famous BERT model (Devlin et al. 2019), synthesizing over 40 analysis studies. We also provide an overview of the proposed modifications to the model and its training regime. We then outline the directions for further research."]]></description>
<dc:subject>neural_networks natural_language_processing text_mining to_read large_language_models_(so_called) in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:77a75c67f585/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:large_language_models_(so_called)"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://cloud.ibm.com/docs/personality-insights?topic=personality-insights-science">
    <title>The science behind the service [IBM Watson Personality Insights]</title>
    <dc:date>2020-07-16T18:57:42+00:00</dc:date>
    <link>https://cloud.ibm.com/docs/personality-insights?topic=personality-insights-science</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA[Words on Twitter -> Glove embedding -> prediction of Big 5 personality scores (et al.).

Correlations of about 0.3 for English in their samples where they actually administered personality tests.  No details about how the subjects were recruited, or even whether this is the correlation on a test set or on a training set.


--- ETA: At the bottom of the document, they admit that "While the correlation between inferred and survey-based scores is both positive and significant, the results imply that inferred scores might not always correlate with survey-based results", and give three references to three non-IBM publications which (supposedly) "conducted experiments to compare how well inferred scores match scores obtained from surveys".  The third of these is: "Mairesse and Walker (2006) reported 60- to 70-percent accuracy for Big Five personality prediction."  I was intrigued by what 60% accuracy would mean for continuous vectors, so I followed their link; this reported both regression results and classifications where the threshold between the class was, for each personality dimension, set at the median.  What's remarkable there is that _most_ of the reported results are not marked as statistically significant improvements over the baseline (always predict the mean or always predict the more common class).  For self-reported speech, they do 40 hypothesis test and get 2 (!) significant improvements over baseline results at the 5% level.  For corpora judged by others, the models do better; over-all I count 130 hypothesis tests and 36 results significant at the 5% level, so it's not _quite_ a case for the neutral model of inquiry, but it's Pluto's Republic all the way down.]]></description>
<dc:subject>personality_tests text_mining data_mining trapped_in_plutos_republic to:blog re:career_advising_in_plutos_republic</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:643a68d73f09/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:personality_tests"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:data_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:trapped_in_plutos_republic"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:blog"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:career_advising_in_plutos_republic"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.pnas.org/content/116/52/26459">
    <title>Social media-predicted personality traits and values can help match people to their ideal jobs | PNAS</title>
    <dc:date>2020-07-16T15:49:42+00:00</dc:date>
    <link>https://www.pnas.org/content/116/52/26459</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Work is thought to be more enjoyable and beneficial to individuals and society when there is congruence between one’s personality and one’s occupation. We provide large-scale evidence that occupations have distinctive psychological profiles, which can successfully be predicted from linguistic information unobtrusively collected through social media. Based on 128,279 Twitter users representing 3,513 occupations, we automatically assess user personalities and visually map the personality profiles of different professions. Similar occupations cluster together, pointing to specific sets of jobs that one might be well suited for. Observations that contradict existing classifications may point to emerging occupations relevant to the 21st century workplace. Findings illustrate how social media can be used to match people to their ideal occupation."

--- Some observations:
1. They did not actually measure people's personality traits; they _assumed_ that a commercial IBM product can map word usage to personality traits.
1a. In particular, they _assumed_ that this remains accurate for what people write on Twitter, as opposed to whatever context IBM developed their system in (not specified here).
2. They did not actually measure "ideal" occupations; they saw whether a classifier using the estimated personality traits could map people to their actual occupations.
2a. They artificially balance their 10 professions so that each has 955 members.  (I presume that they randomly sampled the occupations with more members, though I don't quite see them saying that; maybe I missed it.  Also, I presume they did _not_ go hunting for the best group of 10 occupations.)  So the baseline accuracy would be only 10%, and getting about 70% under CV does indeed mean that there's some signal here.
2b. It's good that they include error bars on their accuracy figures!
2c.  Since they include those error bars, we can see that the difference in classification accuracy between the different methods are both small and statistically insignificant.  In particular, good old fashioned logistic regression is pretty much on par with everything else.
2d. They don't seem to have actually tried the obvious classifier here, which would map each person to the occupation whose feature-vector center ("medoid") was closest to the person's feature-vector ("prototype method").  But they did at least use k-nearest-neighbors, which performed about as well as all the others.
3. Calling this evidence that we could go from analyzing Twitter word usage to "ideal" job recommendations presumes that most people are _already_ in their ideal jobs.
4. This was edited by Susan Fiske [https://statmodeling.stat.columbia.edu/2017/02/08/authority-figures-spread-happy-talk-still-dont-get-it/].

_Maybe_ people reveal their personalities, in the Big 5 sense, by what they write on Twitter.  (Operationally, "personality" in the Big 5 sense is pretty close to "what words would you use to describe yourself on a questionnaire?")  And _maybe_ the way people reveal their personalities in their word usage on Twitter is so context-independent that it can reliably generalize across all the different sub-cultures and sub-societies and self-organized genre conventions of Twitter, so there is one globally reliable mapping.  (I am not going to repeat all of [http://bactra.org/weblog/770.html], but I could.)  And _maybe_ IBM has provided that mapping with an API.  And _maybe_ people with different personalities select in to different professions.  (As an alternative: different occupations train people differently, which alters their personalities, or at least the verbal expressions thereof, and different occupations expose people to different situations, which alters what they say and maybe even shapes their personalities.)  And _maybe_ people select in to professions where they are happier.  And _maybe_ if we looked at how young people talk on Twitter, before they've chosen an occupation, and extract their personality from it, and map them to a profession with lots of similar personality vectors already in it, they'll be happier in that occupation than in others.  But this study provides at best very, very weak evidence for all this.  (I want to say "no evidence at all", but I also don't want to get into arguments about the theory of evidence.)  What the study does show is that people in different occupations use different words on Twitter, and that these differences are detectable through the filter of IBM's purported personality estimator.

]]></description>
<dc:subject>to:NB have_read bad_science bad_data_analysis classifiers text_mining personality_tests logistic_regression social_media psychology why_oh_why_cant_we_have_a_better_academic_publishing_system to_teach:data-mining forty_minutes_of_my_life_im_not_getting_back trapped_in_plutos_republic to:blog twitter re:career_advising_in_plutos_republic</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:780cca65f6d0/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_science"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:bad_data_analysis"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:classifiers"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:personality_tests"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:logistic_regression"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:social_media"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:psychology"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:why_oh_why_cant_we_have_a_better_academic_publishing_system"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:forty_minutes_of_my_life_im_not_getting_back"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:trapped_in_plutos_republic"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:blog"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:twitter"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:career_advising_in_plutos_republic"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1905.12516">
    <title>[1905.12516] Racial Bias in Hate Speech and Abusive Language Detection Datasets</title>
    <dc:date>2020-05-06T20:12:10+00:00</dc:date>
    <link>https://arxiv.org/abs/1905.12516</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Technologies for abusive language detection are being developed and applied with little consideration of their potential biases. We examine racial bias in five different sets of Twitter data annotated for hate speech and abusive language. We train classifiers on these datasets and compare the predictions of these classifiers on tweets written in African-American English with those written in Standard American English. The results show evidence of systematic racial bias in all datasets, as classifiers trained on them tend to predict that tweets written in African-American English are abusive at substantially higher rates. If these abusive language detection systems are used in the field they will therefore have a disproportionate negative impact on African-American social media users. Consequently, these systems may discriminate against the groups who are often the targets of the abuse we are trying to detect."]]></description>
<dc:subject>algorithmic_fairness text_mining classifiers to_teach:data-mining to_teach:statistics_of_inequality_and_discrimination in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:1ce9a1ebca05/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:algorithmic_fairness"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:classifiers"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:statistics_of_inequality_and_discrimination"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.jair.org/index.php/jair/article/view/11030">
    <title>A Primer on Neural Network Models for Natural Language Processing | Journal of Artificial Intelligence Research</title>
    <dc:date>2019-11-25T15:57:00+00:00</dc:date>
    <link>https://www.jair.org/index.php/jair/article/view/11030</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Over the past few years, neural networks have re-emerged as powerful machine-learning models, yielding state-of-the-art results in fields such as image recognition and speech processing. More recently, neural network models started to be applied also to textual natural language signals, again with very promising results. This tutorial surveys neural network models from the perspective of natural language processing research, in an attempt to bring natural-language researchers up to speed with the neural techniques. The tutorial covers input encoding for natural language tasks, feed-forward networks, convolutional networks, recurrent networks and recursive networks, as well as the computation graph abstraction for automatic gradient computation."]]></description>
<dc:subject>to:NB neural_networks natural_language_processing text_mining</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:519570246914/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1911.02656">
    <title>[1911.02656] Invariance and identifiability issues for word embeddings</title>
    <dc:date>2019-11-11T20:13:17+00:00</dc:date>
    <link>https://arxiv.org/abs/1911.02656</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Word embeddings are commonly obtained as optimizers of a criterion function f of a text corpus, but assessed on word-task performance using a different evaluation function g of the test data. We contend that a possible source of disparity in performance on tasks is the incompatibility between classes of transformations that leave f and g invariant. In particular, word embeddings defined by f are not unique; they are defined only up to a class of transformations to which f is invariant, and this class is larger than the class to which g is invariant. One implication of this is that the apparent superiority of one word embedding over another, as measured by word task performance, may largely be a consequence of the arbitrary elements selected from the respective solution sets. We provide a formal treatment of the above identifiability issue, present some numerical examples, and discuss possible resolutions."]]></description>
<dc:subject>word_embeddings text_mining natural_language_processing model_selection to_teach:data-mining have_read linear_algebra oopsies in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:15724565eb13/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:word_embeddings"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:model_selection"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:linear_algebra"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:oopsies"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1911.02639">
    <title>[1911.02639] Word Embedding Algorithms as Generalized Low Rank Models and their Canonical Form</title>
    <dc:date>2019-11-11T20:10:33+00:00</dc:date>
    <link>https://arxiv.org/abs/1911.02639</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Word embedding algorithms produce very reliable feature representations of words that are used by neural network models across a constantly growing multitude of NLP tasks. As such, it is imperative for NLP practitioners to understand how their word representations are produced, and why they are so impactful.
"The present work presents the Simple Embedder framework, generalizing the state-of-the-art existing word embedding algorithms (including Word2vec (SGNS) and GloVe) under the umbrella of generalized low rank models. We derive that both of these algorithms attempt to produce embedding inner products that approximate pointwise mutual information (PMI) statistics in the corpus. Once cast as Simple Embedders, comparison of these models reveals that these successful embedders all resemble a straightforward maximum likelihood estimate (MLE) of the PMI parametrized by the inner product (between embeddings). This MLE induces our proposed novel word embedding model, Hilbert-MLE, as the canonical representative of the Simple Embedder framework.
"We empirically compare these algorithms with evaluations on 17 different datasets. Hilbert-MLE consistently observes second-best performance on every extrinsic evaluation (news classification, sentiment analysis, POS-tagging, and supersense tagging), while the first-best model depends varying on the task. Moreover, Hilbert-MLE consistently observes the least variance in results with respect to the random initialization of the weights in bidirectional LSTMs. Our empirical results demonstrate that Hilbert-MLE is a very consistent word embedding algorithm that can be reliably integrated into existing NLP systems to obtain high-quality results."]]></description>
<dc:subject>text_mining natural_language_processing word_embeddings information_theory to_teach:data-mining low-rank_approximation have_read in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:2ea2e7a59cfb/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:word_embeddings"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:information_theory"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:low-rank_approximation"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1910.12618">
    <title>[1910.12618] Textual Data for Time Series Forecasting</title>
    <dc:date>2019-10-29T14:59:54+00:00</dc:date>
    <link>https://arxiv.org/abs/1910.12618</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["While ubiquitous, textual sources of information such as company reports, social media posts, etc. are hardly included in prediction algorithms for time series, despite the relevant information they may contain. In this work, openly accessible daily weather reports from France and the United-Kingdom are leveraged to predict time series of national electricity consumption, average temperature and wind-speed with a single pipeline. Two methods of numerical representation of text are considered, namely traditional Term Frequency - Inverse Document Frequency (TF-IDF) as well as our own neural word embedding. Using exclusively text, we are able to predict the aforementioned time series with sufficient accuracy to be used to replace missing data. Furthermore the proposed word embeddings display geometric properties relating to the behavior of the time series and context similarity between words."]]></description>
<dc:subject>to:NB text_mining natural_language_processing prediction time_series statistics</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:ac01a1a4a360/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:prediction"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:time_series"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:statistics"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1910.12203">
    <title>[1910.12203] Do Sentence Interactions Matter? Leveraging Sentence Level Representations for Fake News Classification</title>
    <dc:date>2019-10-29T14:15:52+00:00</dc:date>
    <link>https://arxiv.org/abs/1910.12203</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["The rising growth of fake news and misleading information through online media outlets demands an automatic method for detecting such news articles. Of the few limited works which differentiate between trusted vs other types of news article (satire, propaganda, hoax), none of them model sentence interactions within a document. We observe an interesting pattern in the way sentences interact with each other across different kind of news articles. To capture this kind of information for long news articles, we propose a graph neural network-based model which does away with the need of feature engineering for fine grained fake news classification. Through experiments, we show that our proposed method beats strong neural baselines and achieves state-of-the-art accuracy on existing datasets. Moreover, we establish the generalizability of our model by evaluating its performance in out-of-domain scenarios. Code is available at this https URL"]]></description>
<dc:subject>to:NB text_mining deceiving_us_has_become_an_industrial_process networked_life re:disinformation_attacks_on_democratic_systems</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:c345a4d890eb/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:deceiving_us_has_become_an_industrial_process"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:networked_life"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:disinformation_attacks_on_democratic_systems"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1910.12073">
    <title>[1910.12073] Disinformation Detection: A review of linguistic feature selection and classification models in news veracity assessments</title>
    <dc:date>2019-10-29T14:15:00+00:00</dc:date>
    <link>https://arxiv.org/abs/1910.12073</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Over the past couple of years, the topic of "fake news" and its influence over people's opinions has become a growing cause for concern. Although the spread of disinformation on the Internet is not a new phenomenon, the widespread use of social media has exacerbated its effects, providing more channels for dissemination and the potential to "go viral." Nowhere was this more evident than during the 2016 United States Presidential Election. Although the current of disinformation spread via trolls, bots, and hyperpartisan media outlets likely reinforced existing biases rather than sway undecided voters, the effects of this deluge of disinformation are by no means trivial. The consequences range in severity from an overall distrust in news media, to an ill-informed citizenry, and in extreme cases, provocation of violent action. It is clear that human ability to discern lies from truth is flawed at best. As such, greater attention has been given towards applying machine learning approaches to detect deliberately deceptive news articles. This paper looks at the work that has already been done in this area."]]></description>
<dc:subject>to:NB networked_life text_mining re:disinformation_attacks_on_democratic_systems</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:1ba9c6c15ff4/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:networked_life"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:re:disinformation_attacks_on_democratic_systems"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1910.08350">
    <title>[1910.08350] A Mutual Information Maximization Perspective of Language Representation Learning</title>
    <dc:date>2019-10-24T13:41:05+00:00</dc:date>
    <link>https://arxiv.org/abs/1910.08350</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["We show state-of-the-art word representation learning methods maximize an objective function that is a lower bound on the mutual information between different parts of a word sequence (i.e., a sentence). Our formulation provides an alternative perspective that unifies classical word embedding models (e.g., Skip-gram) and modern contextual embeddings (e.g., BERT, XLNet). In addition to enhancing our theoretical understanding of these methods, our derivation leads to a principled framework that can be used to construct new self-supervised tasks. We provide an example by drawing inspirations from related methods based on mutual information maximization that have been successful in computer vision, and introduce a simple self-supervised objective that maximizes the mutual information between a global sentence representation and n-grams in the sentence. Our analysis offers a holistic view of representation learning methods to transfer knowledge and translate progress across multiple domains (e.g., natural language processing, computer vision, audio processing)."

--- This would have been very useful to read _before_ explaining word2vec et al. to The Kids yesterday.]]></description>
<dc:subject>information_theory natural_language_processing text_mining to_teach:data-mining have_read in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:5c06199bbf45/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:information_theory"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to_teach:data-mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:have_read"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1907.04670">
    <title>[1907.04670] Comparing the Performance of the LSTM and HMM Language Models via Structural Similarity</title>
    <dc:date>2019-10-17T14:13:01+00:00</dc:date>
    <link>https://arxiv.org/abs/1907.04670</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Language models based on deep neural networks and traditional stochastic modelling have become both highly functional and effective in recent times. In this work, a general survey into the two types of language modelling is conducted. We investigate the effectiveness of the Hidden Markov Model (HMM), and the Long Short-Term Memory Model (LSTM). We analyze the hidden state structures common to both models, and present an analysis on structural similarity of the hidden states, common to both HMM's and LSTM's. We compare the LSTM's predictive accuracy and hidden state output with respect to the HMM for a varying number of hidden states. In this work, we justify that the less complex HMM can serve as an appropriate approximation of the LSTM model."]]></description>
<dc:subject>to:NB natural_language_processing text_mining markov_models neural_networks statistics your_favorite_deep_neural_network_sucks</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:e02f8d1668fc/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:natural_language_processing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:markov_models"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:statistics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:your_favorite_deep_neural_network_sucks"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://languagelog.ldc.upenn.edu/nll/?p=44621">
    <title>Language Log » Danger: Demo!</title>
    <dc:date>2019-10-11T00:39:58+00:00</dc:date>
    <link>https://languagelog.ldc.upenn.edu/nll/?p=44621</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["And Seabrook's experience illustrates a crucial lesson that AI researchers learned 40 or 50 years ago: "evaluation by demonstration" is a recipe for what John Pierce called glamor and (self-) deceit ("Whither Speech Recognition", JASA 1969). Why? Because we humans are prone to over-generalizing and anthropomorphizing the behavior of machines; and because someone who wants to show how good a system is will choose successful examples and discard failures. I'd be surprised if Seabrook didn't do a bit of this in creating and selecting his "Read Predicted Text" examples."]]></description>
<dc:subject>to:NB text_mining neural_networks machine_learning artificial_intelligence liberman.mark</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:adfa2b0c27f1/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:neural_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:artificial_intelligence"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:liberman.mark"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1910.00163">
    <title>[1910.00163] Specializing Word Embeddings (for Parsing) by Information Bottleneck</title>
    <dc:date>2019-10-02T15:53:00+00:00</dc:date>
    <link>https://arxiv.org/abs/1910.00163</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["Pre-trained word embeddings like ELMo and BERT contain rich syntactic and semantic information, resulting in state-of-the-art performance on various tasks. We propose a very fast variational information bottleneck (VIB) method to nonlinearly compress these embeddings, keeping only the information that helps a discriminative parser. We compress each word embedding to either a discrete tag or a continuous vector. In the discrete version, our automatically compressed tags form an alternative tag set: we show experimentally that our tags capture most of the information in traditional POS tag annotations, but our tag sequences can be parsed more accurately at the same level of tag granularity. In the continuous version, we show experimentally that moderately compressing the word embeddings by our method yields a more accurate parser in 8 of 9 languages, unlike simple dimensionality reduction."]]></description>
<dc:subject>to:NB information_bottleneck text_mining inference_to_latent_objects statistics</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:55644087a65d/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:to:NB"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:information_bottleneck"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:inference_to_latent_objects"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:statistics"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1909.13456">
    <title>[1909.13456] Improving Textual Network Learning with Variational Homophilic Embeddings</title>
    <dc:date>2019-10-01T17:15:12+00:00</dc:date>
    <link>https://arxiv.org/abs/1909.13456</link>
    <dc:creator>cshalizi</dc:creator><description><![CDATA["The performance of many network learning applications crucially hinges on the success of network embedding algorithms, which aim to encode rich network information into low-dimensional vertex-based vector representations. This paper considers a novel variational formulation of network embeddings, with special focus on textual networks. Different from most existing methods that optimize a discriminative objective, we introduce Variational Homophilic Embedding (VHE), a fully generative model that learns network embeddings by modeling the semantic (textual) information with a variational autoencoder, while accounting for the structural (topology) information through a novel homophilic prior design. Homophilic vertex embeddings encourage similar embedding vectors for related (connected) vertices. The proposed VHE promises better generalization for downstream tasks, robustness to incomplete observations, and the ability to generalize to unseen vertices. Extensive experiments on real-world networks, for multiple tasks, demonstrate that the proposed method consistently achieves superior performance relative to competing state-of-the-art approaches."]]></description>
<dc:subject>text_mining network_data_analysis joint_modeling_of_text_and_networks in_NB</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:cshalizi/b:4a3815969b5c/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:text_mining"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:network_data_analysis"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:joint_modeling_of_text_and_networks"/>
	<rdf:li rdf:resource="https://pinboard.in/u:cshalizi/t:in_NB"/>
</rdf:Bag></taxo:topics>
</item>
</rdf:RDF>