<?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 (amy)</title>
    <link>https://pinboard.in/u:amy/public/</link>
    <description>recent bookmarks from amy</description>
    <items>
      <rdf:Seq>	<rdf:li rdf:resource="https://allenai.org/blog/olmoasr"/>
	<rdf:li rdf:resource="https://www.nltk.org/_modules/nltk/sentiment/vader.html"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1902.00751"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2301.04655"/>
	<rdf:li rdf:resource="https://mashable.com/article/chatgpt-amazing-wrong?s=09"/>
	<rdf:li rdf:resource="https://www.nytimes.com/2022/12/05/technology/chatgpt-ai-twitter.html?smid=nytcore-ios-share&amp;referringSource=articleShare"/>
	<rdf:li rdf:resource="https://www.jasonwei.net/blog/emergence"/>
	<rdf:li rdf:resource="https://www.washingtonpost.com/technology/2022/07/21/big-science-ai-open-source-language-model/"/>
	<rdf:li rdf:resource="https://www.technologyreview.com/2022/07/12/1055817/inside-a-radical-new-project-to-democratize-ai/"/>
	<rdf:li rdf:resource="https://openai.com/api/?s=09"/>
	<rdf:li rdf:resource="https://www.businesswire.com/news/home/20220426005963/en/AI-Transformer-Inventors-Launch-Adept-with-65M-to-Lend-a-Hand-to-Knowledge-Workers"/>
	<rdf:li rdf:resource="https://www.marktechpost.com/2022/04/04/google-ais-latest-540-billion-parameter-model-pathways-language-model-called-palm-unlocks-new-tasks-proportional-to-scale/"/>
	<rdf:li rdf:resource="https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html"/>
	<rdf:li rdf:resource="https://cloud.google.com/blog/topics/public-datasets/data-and-ai-google-cloud-datasets-and-new-data-benchmarks"/>
	<rdf:li rdf:resource="https://huggingface.co/docs/transformers/model_doc/clip"/>
	<rdf:li rdf:resource="https://www.npmjs.com/package/@tensorflow-models/universal-sentence-encoder"/>
	<rdf:li rdf:resource="https://workspace.google.com/marketplace/app/semantic_reactor/509042860915"/>
	<rdf:li rdf:resource="https://github.com/recognai/rubrix"/>
	<rdf:li rdf:resource="https://simpletransformers.ai/about/"/>
	<rdf:li rdf:resource="https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html"/>
	<rdf:li rdf:resource="https://lastweekin.ai/p/gpt-3-foundation-models-and-ai-nationalism"/>
	<rdf:li rdf:resource="https://medium.com/google-cloud/accelerating-question-answering-applications-with-mobilebert-and-tflite-1f8d301ddbf7"/>
	<rdf:li rdf:resource="https://ai.googleblog.com/2021/09/announcing-wit-wikipedia-based-image.html"/>
	<rdf:li rdf:resource="https://aylien.com/blog/leveraging-deep-learning-for-multilingual"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2005.14165v4"/>
	<rdf:li rdf:resource="https://www.fastcompany.com/90666920/ai-bias-stanford-percy-liang-fei-fei-li"/>
	<rdf:li rdf:resource="https://www.fastcompany.com/90670635/ex-googlers-raise-40-million-to-democratize-natural-language-ai"/>
	<rdf:li rdf:resource="https://faculty.washington.edu/ebender/2021_575/"/>
	<rdf:li rdf:resource="https://ai.googleblog.com/2020/10/rethinking-attention-with-performers.html"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2108.07258"/>
	<rdf:li rdf:resource="https://github.com/google-research/bert"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2106.16171"/>
	<rdf:li rdf:resource="https://www.youtube.com/watch?v=LE3NfEULV6k"/>
	<rdf:li rdf:resource="https://github.com/huggingface/transformers/tree/master/examples/flax/text-classification"/>
	<rdf:li rdf:resource="https://github.com/deepset-ai/haystack"/>
	<rdf:li rdf:resource="https://www.marktechpost.com/2021/07/13/stanfords-ai-researchers-introduce-qa-gnn-that-jointly-reasons-with-language-models-and-knowledge-graphs/"/>
	<rdf:li rdf:resource="https://github.com/dimitreOliveira/bert-as-a-service_TFX"/>
	<rdf:li rdf:resource="https://bit.ly/3bVF0q3"/>
	<rdf:li rdf:resource="https://blog.google/products/search/introducing-mum/"/>
	<rdf:li rdf:resource="https://www.wired.com/story/ai-generate-convincing-text-anyone-use-it/?mc_cid=90e7741fa8&amp;mc_eid=eef4b324fd"/>
	<rdf:li rdf:resource="https://arxiv.org/pdf/2101.03961.pdf"/>
	<rdf:li rdf:resource="https://venturebeat.com/2021/02/09/openai-and-stanford-researchers-call-for-urgent-action-to-address-harms-of-large-language-models-like-gpt-3/"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2002.09402"/>
	<rdf:li rdf:resource="https://openai.com/blog/dall-e/"/>
	<rdf:li rdf:resource="https://twitter.com/nabla_theta/status/1345130408170541056"/>
	<rdf:li rdf:resource="https://github.com/google/patents-public-data/blob/master/models/BERT%20for%20Patents.md"/>
	<rdf:li rdf:resource="https://cloud.google.com/blog/products/ai-machine-learning/how-ai-improves-patent-analysis"/>
	<rdf:li rdf:resource="https://github.com/NiuTrans/ABigSurvey"/>
	<rdf:li rdf:resource="https://gradientflow.com/2020nlpsurvey/"/>
	<rdf:li rdf:resource="https://bdtechtalks.com/2020/08/17/openai-gpt-3-commercial-ai/amp/"/>
	<rdf:li rdf:resource="https://github.com/google/making_with_ml/blob/master/semantic_ml/use_sample.js"/>
	<rdf:li rdf:resource="https://blog.tensorflow.org/2020/08/introducing-semantic-reactor-explore-nlp-sheets.html?m=1"/>
	<rdf:li rdf:resource="https://ai.googleblog.com/2018/05/advances-in-semantic-textual-similarity.html?m=1"/>
	<rdf:li rdf:resource="https://www.technologyreview.com/2020/08/14/1006780/ai-gpt-3-fake-blog-reached-top-of-hacker-news"/>
	<rdf:li rdf:resource="https://www.sudowrite.com/"/>
	<rdf:li rdf:resource="https://ai.googleblog.com/2020/04/a-scalable-approach-to-reducing-gender.html"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2002.08909"/>
	<rdf:li rdf:resource="https://towardsdatascience.com/gpt-3-creative-potential-of-nlp-d5ccae16c1ab"/>
	<rdf:li rdf:resource="https://drive.google.com/file/d/19NicWkWcF_-V0UvxUqAzNoRbT-Dp6SPK/view"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/2004.15011"/>
	<rdf:li rdf:resource="https://scitldr.apps.allenai.org/"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1905.07129v3"/>
	<rdf:li rdf:resource="https://github.com/tensorflow/models/tree/master/official/nlp/xlnet"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1906.08237v2"/>
	<rdf:li rdf:resource="https://www.analyticsvidhya.com/blog/2018/03/essentials-of-deep-learning-sequence-to-sequence-modelling-with-attention-part-i/"/>
	<rdf:li rdf:resource="https://www.analyticsvidhya.com/blog/2019/06/understanding-transformers-nlp-state-of-the-art-models/"/>
	<rdf:li rdf:resource="https://github.com/zihangdai/xlnet"/>
	<rdf:li rdf:resource="https://www.analyticsvidhya.com/blog/2020/03/6-pretrained-models-text-classification/"/>
	<rdf:li rdf:resource="https://arxiv.org/abs/1910.10683"/>
	<rdf:li rdf:resource="https://github.com/google-research/text-to-text-transfer-transformer"/>
      </rdf:Seq>
    </items>
  </channel><item rdf:about="https://allenai.org/blog/olmoasr">
    <title>OLMoASR: A series of open speech recognition models | Ai2</title>
    <dc:date>2025-09-01T14:47:38+00:00</dc:date>
    <link>https://allenai.org/blog/olmoasr</link>
    <dc:creator>amy</dc:creator><description><![CDATA[We release OLMoASR, a family of open automatic speech recognition (ASR) models trained from scratch on a curated, large-scale dataset.]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:b01681bdd687/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.nltk.org/_modules/nltk/sentiment/vader.html">
    <title>nltk.sentiment.vader</title>
    <dc:date>2024-06-02T11:51:27+00:00</dc:date>
    <link>https://www.nltk.org/_modules/nltk/sentiment/vader.html</link>
    <dc:creator>amy</dc:creator><dc:subject>nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:61969ff3c9c3/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1902.00751">
    <title>[1902.00751] Parameter-Efficient Transfer Learning for NLP</title>
    <dc:date>2023-01-26T22:50:29+00:00</dc:date>
    <link>https://arxiv.org/abs/1902.00751</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Fine-tuning large pre-trained models is an effective transfer mechanism in NLP. However, in the presence of many downstream tasks, fine-tuning is parameter inefficient: an entire new model is required for every task. As an alternative, we propose transfer with adapter modules. Adapter modules yield a compact and extensible model; they add only a few trainable parameters per task, and new tasks can be added without revisiting previous ones. The parameters of the original network remain fixed, yielding a high degree of parameter sharing. To demonstrate adapter's effectiveness, we transfer the recently proposed BERT Transformer model to 26 diverse text classification tasks, including the GLUE benchmark. Adapters attain near state-of-the-art performance, whilst adding only a few parameters per task. On GLUE, we attain within 0.4% of the performance of full fine-tuning, adding only 3.6% parameters per task. By contrast, fine-tuning trains 100% of the parameters per task.]]></description>
<dc:subject>nlp google machine_learning</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:1c8a13a4e72b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2301.04655">
    <title>[2301.04655] ChatGPT is not all you need. A State of the Art Review of large Generative AI models</title>
    <dc:date>2023-01-19T16:40:30+00:00</dc:date>
    <link>https://arxiv.org/abs/2301.04655</link>
    <dc:creator>amy</dc:creator><description><![CDATA[During the last two years there has been a plethora of large generative models such as ChatGPT or Stable Diffusion that have been published. Concretely, these models are able to perform tasks such as being a general question and answering system or automatically creating artistic images that are revolutionizing several sectors. Consequently, the implications that these generative models have in the industry and society are enormous, as several job positions may be transformed. For example, Generative AI is capable of transforming effectively and creatively texts to images, like the DALLE-2 model; text to 3D images, like the Dreamfusion model; images to text, like the Flamingo model; texts to video, like the Phenaki model; texts to audio, like the AudioLM model; texts to other texts, like ChatGPT; texts to code, like the Codex model; texts to scientific texts, like the Galactica model or even create algorithms like AlphaTensor. This work consists on an attempt to describe in a concise way the main models are sectors that are affected by generative AI and to provide a taxonomy of the main generative models published recently.]]></description>
<dc:subject>machine_learning nlp llms</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:3bb4039e2475/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:llms"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://mashable.com/article/chatgpt-amazing-wrong?s=09">
    <title>ChatGPT from OpenAI is a huge step toward a usable answer engine. Unfortunately its answers are horrible. | Mashable</title>
    <dc:date>2022-12-08T15:12:24+00:00</dc:date>
    <link>https://mashable.com/article/chatgpt-amazing-wrong?s=09</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:1b086e2397f3/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.nytimes.com/2022/12/05/technology/chatgpt-ai-twitter.html?smid=nytcore-ios-share&amp;referringSource=articleShare">
    <title>The Brilliance and Weirdness of ChatGPT</title>
    <dc:date>2022-12-07T01:45:28+00:00</dc:date>
    <link>https://www.nytimes.com/2022/12/05/technology/chatgpt-ai-twitter.html?smid=nytcore-ios-share&amp;referringSource=articleShare</link>
    <dc:creator>amy</dc:creator><description><![CDATA[A new chatbot from OpenAI is inspiring awe, fear, stunts and attempts to circumvent its guardrails.]]></description>
<dc:subject>machine_learning NLP</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:f1e3f8fdc422/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.jasonwei.net/blog/emergence">
    <title>137 emergent abilities of large language models — Jason Wei</title>
    <dc:date>2022-11-29T14:33:21+00:00</dc:date>
    <link>https://www.jasonwei.net/blog/emergence</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:1e99942862a6/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.washingtonpost.com/technology/2022/07/21/big-science-ai-open-source-language-model/">
    <title>Big Tech builds AI with bad data. So scientists sought better data.</title>
    <dc:date>2022-07-23T04:01:10+00:00</dc:date>
    <link>https://www.washingtonpost.com/technology/2022/07/21/big-science-ai-open-source-language-model/</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:b74c825a7a40/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.technologyreview.com/2022/07/12/1055817/inside-a-radical-new-project-to-democratize-ai/">
    <title>BLOOM: Inside the radical new project to democratize AI | MIT Technology Review</title>
    <dc:date>2022-07-14T03:00:19+00:00</dc:date>
    <link>https://www.technologyreview.com/2022/07/12/1055817/inside-a-radical-new-project-to-democratize-ai/</link>
    <dc:creator>amy</dc:creator><description><![CDATA[A group of over 1,000 AI researchers has created a multilingual large language model bigger than GPT-3—and they’re giving it out for free.]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:9b1c12801bf7/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://openai.com/api/?s=09">
    <title>OpenAI API</title>
    <dc:date>2022-07-02T02:45:59+00:00</dc:date>
    <link>https://openai.com/api/?s=09</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:f1a6ad784815/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.businesswire.com/news/home/20220426005963/en/AI-Transformer-Inventors-Launch-Adept-with-65M-to-Lend-a-Hand-to-Knowledge-Workers">
    <title>AI Transformer Inventors Launch Adept with $65M to Lend a Hand to Knowledge Workers | Business Wire</title>
    <dc:date>2022-04-28T00:13:52+00:00</dc:date>
    <link>https://www.businesswire.com/news/home/20220426005963/en/AI-Transformer-Inventors-Launch-Adept-with-65M-to-Lend-a-Hand-to-Knowledge-Workers</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Adept was founded by a team with deep roots in large-scale neural networks. David Luan (CEO) led Google’s large model program, and before that was head of engineering at OpenAI. Ashish Vaswani (Chief Scientist) and Niki Parmar (CTO) are former Google Brain researchers and inventors of the Transformer, a model architecture that dramatically improved the way computers understand natural language. Since their invention in 2017, Transformers have unlocked decades-old problems in rapid succession – powering groundbreaking applications like GPT-3, a model that can generate human-like language such as tweets, emails, and trivia questions.

While at Google, Ashish, Niki, and David trained bigger and bigger Transformers, with the goal of eventually building a model that could power all ML use cases. Along the way, the team discovered a major limitation: models like GPT-3 can write great prose, but they can’t take actions in the digital world.

“Transformers and their applications represent the single largest step towards general intelligence in recent history. But we believe true general intelligence requires Transformers that can act – not just read and write,” said David Luan, CEO and Co-founder of Adept. “At Adept, we’re training a model to use every software tool and API that people use today. ”

Users will work hand-in-hand with Adept’s technology, using a natural language interface to use existing software like Airtable, Photoshop, an ATS, Tableau, and Twilio. Adept sees this product vision as a practical path to general intelligence that puts people in the driver’s seat, enabling people to offload manual tasks and free up time for the work they most enjoy.

“Machine learning has undergone significant advances in the last few years since the Transformer came out in 2017, and Adept’s team has literally invented it along with several other breakthroughs. There are few startups with the same level of technical capability in AI,” said Saam Motamedi, partner at Greylock and board member at Adept. “We’re excited to see Adept drive the next wave of advances through their integrated approach to research and product – pushing the frontier of generally capable models while creating tremendous user and business value.”

]]></description>
<dc:subject>google machine_learning nlp startups</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:a407156c290d/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:startups"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.marktechpost.com/2022/04/04/google-ais-latest-540-billion-parameter-model-pathways-language-model-called-palm-unlocks-new-tasks-proportional-to-scale/">
    <title>Google AI's Latest 540-Billion Parameter Model (Pathways Language Model Called PaLM) Unlocks New Tasks Proportional To Scale - MarkTechPost</title>
    <dc:date>2022-04-06T02:38:09+00:00</dc:date>
    <link>https://www.marktechpost.com/2022/04/04/google-ais-latest-540-billion-parameter-model-pathways-language-model-called-palm-unlocks-new-tasks-proportional-to-scale/</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning google nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:2c9dc969e02d/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html">
    <title>Google AI Blog: Pathways Language Model (PaLM): Scaling to 540 Billion Parameters for Breakthrough Performance</title>
    <dc:date>2022-04-06T01:49:55+00:00</dc:date>
    <link>https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning google NLP</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:431300cd8138/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://cloud.google.com/blog/topics/public-datasets/data-and-ai-google-cloud-datasets-and-new-data-benchmarks">
    <title>Data and AI: Google Cloud datasets and new data benchmarks | Google Cloud Blog</title>
    <dc:date>2022-03-30T00:24:49+00:00</dc:date>
    <link>https://cloud.google.com/blog/topics/public-datasets/data-and-ai-google-cloud-datasets-and-new-data-benchmarks</link>
    <dc:creator>amy</dc:creator><description><![CDATA[At Google, we are excited to contribute to data-centric AI. Today, Google Cloud is adding a new high value dataset to the Public Dataset Program, and Google researchers are announcing DataPerf, a new multi-organizational effort to develop benchmarks for data quality and data centric algorithms.

Google Cloud is committed to helping users improve their data quality, starting with supporting better public data. The Public Datasets program provides high quality datasets pre-configured on GCP for easy access. Google Cloud is adding a new high-value dataset developed by the MLCommons™ Association (which Google co-founded) to the Public Datasets program: The Multilingual Spoken Words Corpus: a rich audio speech dataset with more than 340,000 keywords in 50 languages with upwards of 23.4 million examples.

This new public dataset is aligned with the MLCommons Association vision for “open” datasets – accessible by all – that are “living” – continually being improved to raise quality and increase representation and diversity.

Google researchers, in collaboration with multiple organizations, are announcing the DataPerf effort at the NeurIPS Data-Centric AI workshop today, to develop benchmarks to improve data quality. Much like the the MLPerf™ benchmarking effort which is now the industry standard for machine learning hardware/software speed, DataPerf brings together the originators of prior efforts including: CATS4ML, Data-Centric AI Competition, DCBench, Dynabench, and the MLPerf benchmarks to define clear metrics that catalyze rapid innovation. DataPerf will measure the utility of training and test data for common problems, and algorithms for working with datasets such as: selecting core sets, correcting errors, identifying under-optimized data slices, and valuing datasets prior to labeling.]]></description>
<dc:subject>google big_data gcp nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:e8e1fee5f113/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:big_data"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:gcp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://huggingface.co/docs/transformers/model_doc/clip">
    <title>CLIP</title>
    <dc:date>2022-03-29T00:48:12+00:00</dc:date>
    <link>https://huggingface.co/docs/transformers/model_doc/clip</link>
    <dc:creator>amy</dc:creator><description><![CDATA[The CLIP model was proposed in Learning Transferable Visual Models From Natural Language Supervision by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever. CLIP (Contrastive Language-Image Pre-Training) is a neural network trained on a variety of (image, text) pairs. It can be instructed in natural language to predict the most relevant text snippet, given an image, without directly optimizing for the task, similarly to the zero-shot capabilities of GPT-2 and 3.]]></description>
<dc:subject>machine_learning NLP embeddings</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:0176e07d750a/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:embeddings"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.npmjs.com/package/@tensorflow-models/universal-sentence-encoder">
    <title>@tensorflow-models/universal-sentence-encoder - npm</title>
    <dc:date>2022-03-29T00:41:13+00:00</dc:date>
    <link>https://www.npmjs.com/package/@tensorflow-models/universal-sentence-encoder</link>
    <dc:creator>amy</dc:creator><description><![CDATA[​​The Universal Sentence Encoder (Cer et al., 2018) (USE) is a model that encodes text into 512-dimensional ​​embeddings. These embeddings can then be used as inputs to natural language processing tasks such as​​ ​​sentiment classification and textual similarity analysis.​
​

This module is a TensorFlow.js GraphModel converted from the USE lite (module on TFHub), a lightweight version of the original. The lite model is based on the Transformer (Vaswani et al, 2017) architecture, and uses an 8k word piece vocabulary.

In this demo we embed six sentences with the USE, and render their self-similarity scores in a matrix]]></description>
<dc:subject>machine_learning nlp nodejs</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:908e23a3051e/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nodejs"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://workspace.google.com/marketplace/app/semantic_reactor/509042860915">
    <title>Semantic Reactor - Google Workspace Marketplace</title>
    <dc:date>2022-03-29T00:37:26+00:00</dc:date>
    <link>https://workspace.google.com/marketplace/app/semantic_reactor/509042860915</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Experiment with machine learning language models.
]]></description>
<dc:subject>machine_learning nlp google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:e599691065de/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/recognai/rubrix">
    <title>GitHub - recognai/rubrix: ✨ Rubrix, open-source framework for data-centric NLP. Data annotation and monitoring for enterprise NLP</title>
    <dc:date>2022-03-24T02:35:30+00:00</dc:date>
    <link>https://github.com/recognai/rubrix</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Rubrix, open-source framework for data-centric NLP. Data annotation and monitoring for enterprise NLP]]></description>
<dc:subject>machine_learning NLP</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:73bae33d73bf/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://simpletransformers.ai/about/">
    <title>About - Simple Transformers</title>
    <dc:date>2022-03-05T01:12:37+00:00</dc:date>
    <link>https://simpletransformers.ai/about/</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Simple Transformers is a Natural Language Processing (NLP) library designed to simplify the usage of Transformer models without having to compromise on utility. It is built on the amazing work of Hugging Face and their Transformers library.]]></description>
<dc:subject>machine_learning transformers nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:7fdf35b3a323/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:transformers"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html">
    <title>Google AI Blog: Exploring Transfer Learning with T5: the Text-To-Text Transfer Transformer</title>
    <dc:date>2022-03-04T22:02:08+00:00</dc:date>
    <link>https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:7500c0b62d3e/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://lastweekin.ai/p/gpt-3-foundation-models-and-ai-nationalism">
    <title>GPT-3, Foundation Models, and AI Nationalism</title>
    <dc:date>2022-02-06T05:23:59+00:00</dc:date>
    <link>https://lastweekin.ai/p/gpt-3-foundation-models-and-ai-nationalism</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning NLP</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:0ed37d94c1ac/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://medium.com/google-cloud/accelerating-question-answering-applications-with-mobilebert-and-tflite-1f8d301ddbf7">
    <title>Accelerating Question Answering Applications with MobileBERT and TFLite | by Felipe de Pontes Adachi | Google Cloud - Community | Nov, 2021 | Medium</title>
    <dc:date>2021-12-10T22:59:09+00:00</dc:date>
    <link>https://medium.com/google-cloud/accelerating-question-answering-applications-with-mobilebert-and-tflite-1f8d301ddbf7</link>
    <dc:creator>amy</dc:creator><description><![CDATA[In this post, I’d like to share our trajectory towards deploying a BERT-family model for Question Answering in order to solve a particular business problem at WEG — automatically detecting technical deviations during a request for quotation for electric motors. These quotations are often accompanied by technical requirements that must be met, and when there’s a specific excerpt that needs to be commented/clarified, the analyst assembling the offer proposal should highlight these points of attention, or deviations.]]></description>
<dc:subject>nlp BERT machine_learning TensorFlow</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:0335bacd64a0/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:BERT"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:TensorFlow"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://ai.googleblog.com/2021/09/announcing-wit-wikipedia-based-image.html">
    <title>Google AI Blog: Announcing WIT: A Wikipedia-Based Image-Text Dataset</title>
    <dc:date>2021-11-27T17:01:25+00:00</dc:date>
    <link>https://ai.googleblog.com/2021/09/announcing-wit-wikipedia-based-image.html</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Multimodal visio-linguistic models rely on rich datasets in order to model the relationship between images and text. Traditionally, these datasets have been created by either manually captioning images, or crawling the web and extracting the alt-text as the caption. While the former approach tends to result in higher quality data, the intensive manual annotation process limits the amount of data that can be created. On the other hand, the automated extraction approach can lead to bigger datasets, but these require either heuristics and careful filtering to ensure data quality or scaling-up models to achieve strong performance. An additional shortcoming of existing datasets is the dearth of coverage in non-English languages. This naturally led us to ask: Can one overcome these limitations and create a high-quality, large-sized, multilingual dataset with a variety of content?

Today we introduce the Wikipedia-Based Image Text (WIT) Dataset, a large multimodal dataset, created by extracting multiple different text selections associated with an image from Wikipedia articles and Wikimedia image links. This was accompanied by rigorous filtering to only retain high quality image-text sets. As detailed in “WIT: Wikipedia-based Image Text Dataset for Multimodal Multilingual Machine Learning”, presented at SIGIR ‘21, this resulted in a curated set of 37.5 million entity-rich image-text examples with 11.5 million unique images across 108 languages. The WIT dataset is available for download and use under the Creative Commons license. We are also excited to announce that we are hosting a competition with the WIT dataset in Kaggle in collaboration with Wikimedia Research and other external collaborators]]></description>
<dc:subject>machine_learning google nlp big_data</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:b21abec18a5d/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:big_data"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://aylien.com/blog/leveraging-deep-learning-for-multilingual">
    <title>Leveraging Deep Learning for Multilingual Sentiment Analysis - AYLIEN News API</title>
    <dc:date>2021-11-23T18:14:33+00:00</dc:date>
    <link>https://aylien.com/blog/leveraging-deep-learning-for-multilingual</link>
    <dc:creator>amy</dc:creator><description><![CDATA[It is a strong indicator of today’s globalized world and rapidly growing access to Internet platforms, that we have users from over 188 countries and 500 cities globally using our Text Analysis and News APIs. Our users need to be able to understand and analyze what’s being said out there, about them, their products, services, or their competitors, regardless of the locality and the language used. Social media content on platforms like Twitter, Facebook and Instagram can provide unrivalled insights into customer opinion and experience to brands and organizations. However, as shown by the following stats, users post content in a multitude of languages on these platforms:

Only about 39% of tweets posted are in English;
Facebook recently reported that about 50% of its users speak a language other than English;
Native platforms such as Sina Weibo and WeChat, where most of the content is written in a native language, are on the rise;
70% of active Instagram users are based outside the US.
A look at online review platforms such as Yelp and TripAdvisor, as well as various news outlets and blogs, reveals similar patterns regarding the variety of language used. Therefore, no matter if you are a social media analyst, or a hotel owner trying to gauge customer satisfaction, or a hedge fund analyst trying to analyze a foreign market, you need to be able to understand textual content in a multitude of languages.]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:ce32f32c27dd/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2005.14165v4">
    <title>(Saving...) [2005.14165v4] Language Models are Few-Shot Learners</title>
    <dc:date>2021-11-23T18:08:06+00:00</dc:date>
    <link>https://arxiv.org/abs/2005.14165v4</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from only a few examples or from simple instructions - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting. For all tasks, GPT-3 is applied without any gradient updates or fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic. At the same time, we also identify some datasets where GPT-3's few-shot learning still struggles, as well as some datasets where GPT-3 faces methodological issues related to training on large web corpora. Finally, we find that GPT-3 can generate samples of news articles which human evaluators have difficulty distinguishing from articles written by humans. We discuss broader societal impacts of this finding and of GPT-3 in general.]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:bc773b55e6f4/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.fastcompany.com/90666920/ai-bias-stanford-percy-liang-fei-fei-li">
    <title>Stanford AI experts warn of biases in GPT-3 and BERT models</title>
    <dc:date>2021-10-07T15:18:44+00:00</dc:date>
    <link>https://www.fastcompany.com/90666920/ai-bias-stanford-percy-liang-fei-fei-li</link>
    <dc:creator>amy</dc:creator><description><![CDATA[A multidisciplinary group of Stanford University professors and students wants to start a serious discussion about the increasing use of large, frighteningly smart, “foundation” AI models such as OpenAI’s GPT-3 (Generative Pretraining Transformer 3) natural language model.

GPT-3 is foundational because it was developed using huge quantities of training data and computer power to reach state-of-the-art, general-purpose performance. Developers, not wanting to reinvent the wheel, are using it as the basis for their software to tackle specific tasks.

But foundation models have some very real downsides, explains Stanford computer science professor Percy Liang. They create “a single point of failure, so any defects, any biases which these models have, any security vulnerabilities . . . are just blindly inherited by all the downstream tasks,” he says.

Liang leads a new group assembled by Stanford’s institute for Human-Centered Artificial Intelligence (HAI) called the Center for Research on Foundation Models (CRFM). The group is studying the impacts and implications of foundation models, and it’s inviting the tech companies developing them to come to the table and participate.]]></description>
<dc:subject>machine_learning nlp ethics</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:27766550aa4e/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:ethics"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.fastcompany.com/90670635/ex-googlers-raise-40-million-to-democratize-natural-language-ai">
    <title>Ex-Googlers raise $40 million to democratize language AI</title>
    <dc:date>2021-10-07T15:16:38+00:00</dc:date>
    <link>https://www.fastcompany.com/90670635/ex-googlers-raise-40-million-to-democratize-natural-language-ai</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:b8e86347861f/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://faculty.washington.edu/ebender/2021_575/">
    <title>Linguistics 575: Societal Impacts of NLP</title>
    <dc:date>2021-10-07T12:37:56+00:00</dc:date>
    <link>https://faculty.washington.edu/ebender/2021_575/</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp education</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:2b3d038ac186/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:education"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://ai.googleblog.com/2020/10/rethinking-attention-with-performers.html">
    <title>Google AI Blog: Rethinking Attention with Performers</title>
    <dc:date>2021-08-23T22:07:22+00:00</dc:date>
    <link>https://ai.googleblog.com/2020/10/rethinking-attention-with-performers.html</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning google nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:9c10cfdb71a2/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2108.07258">
    <title>[2108.07258] On the Opportunities and Risks of Foundation Models</title>
    <dc:date>2021-08-20T01:24:25+00:00</dc:date>
    <link>https://arxiv.org/abs/2108.07258</link>
    <dc:creator>amy</dc:creator><description><![CDATA[AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature.]]></description>
<dc:subject>machine_learning NLP</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:32d32ec063be/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/google-research/bert">
    <title>GitHub - google-research/bert: TensorFlow code and pre-trained models for BERT</title>
    <dc:date>2021-07-30T22:14:30+00:00</dc:date>
    <link>https://github.com/google-research/bert</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning google nlp research</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:18d248ab2bcc/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:research"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2106.16171">
    <title>[2106.16171] Revisiting the Primacy of English in Zero-shot Cross-lingual Transfer</title>
    <dc:date>2021-07-30T22:04:22+00:00</dc:date>
    <link>https://arxiv.org/abs/2106.16171</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Despite their success, large pre-trained multilingual models have not completely alleviated the need for labeled data, which is cumbersome to collect for all target languages. Zero-shot cross-lingual transfer is emerging as a practical solution: pre-trained models later fine-tuned on one transfer language exhibit surprising performance when tested on many target languages. English is the dominant source language for transfer, as reinforced by popular zero-shot benchmarks. However, this default choice has not been systematically vetted. In our study, we compare English against other transfer languages for fine-tuning, on two pre-trained multilingual models (mBERT and mT5) and multiple classification and question answering tasks. We find that other high-resource languages such as German and Russian often transfer more effectively, especially when the set of target languages is diverse or unknown a priori. Unexpectedly, this can be true even when the training sets were automatically translated from English. This finding can have immediate impact on multilingual zero-shot systems, and should inform future benchmark designs.]]></description>
<dc:subject>machine_learning nlp bert t5</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:e8c8341c4ae7/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:bert"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:t5"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.youtube.com/watch?v=LE3NfEULV6k">
    <title>(359) Transfer learning and Transformer models (ML Tech Talks) - YouTube</title>
    <dc:date>2021-07-30T21:24:22+00:00</dc:date>
    <link>https://www.youtube.com/watch?v=LE3NfEULV6k</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp tutorials transformers bert</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:d3ad0d41cedd/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:tutorials"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:transformers"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:bert"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/huggingface/transformers/tree/master/examples/flax/text-classification">
    <title>transformers/examples/flax/text-classification at master · huggingface/transformers</title>
    <dc:date>2021-07-20T16:31:11+00:00</dc:date>
    <link>https://github.com/huggingface/transformers/tree/master/examples/flax/text-classification</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Text classification examples
]]></description>
<dc:subject>nlp jax flax machine_learning huggingface transformers</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:4e08f5ad3256/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:jax"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:flax"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:huggingface"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:transformers"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/deepset-ai/haystack">
    <title>deepset-ai/haystack: End-to-end Python framework for building natural language search interfaces to data. Leverages Transformers and the State-of-the-Art of NLP. Supports DPR, Elasticsearch, Hugging Face’s Hub, and much more!</title>
    <dc:date>2021-07-15T22:26:28+00:00</dc:date>
    <link>https://github.com/deepset-ai/haystack</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Haystack is an end-to-end framework that enables you to build powerful and production-ready pipelines for different search use cases. Whether you want to perform Question Answering or semantic document search, you can use the State-of-the-Art NLP models in Haystack to provide unique search experiences and allow your users to query in natural language. Haystack is built in a modular fashion so that you can combine the best technology from other open-source projects like Huggingface's Transformers, Elasticsearch, or Milvus.]]></description>
<dc:subject>nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:dec4ac2481e3/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.marktechpost.com/2021/07/13/stanfords-ai-researchers-introduce-qa-gnn-that-jointly-reasons-with-language-models-and-knowledge-graphs/">
    <title>Stanford's AI Researchers Introduce QA-GNN Model That Jointly Reasons With Language Models And Knowledge Graphs | MarkTechPost</title>
    <dc:date>2021-07-14T16:58:12+00:00</dc:date>
    <link>https://www.marktechpost.com/2021/07/13/stanfords-ai-researchers-introduce-qa-gnn-that-jointly-reasons-with-language-models-and-knowledge-graphs/</link>
    <dc:creator>amy</dc:creator><description><![CDATA[In recent AI research, background knowledge is usually available in the form of Knowledge Graphs (KGs) and Language Models (LMs) which are pre-trained on a large set of documents. KG’s represent entities as nodes and relations between them as edges, e.g., [Leonardo da Vinci — born in – Italy]. Some other examples of KGs include Freebase (general-purpose facts), ConceptNet (commonsense) and Examples of pre-trained LMs include BERT (trained on Wikipedia articles and 10,000 books), RoBERTa (extending BERT), etc.

Both the knowledge sources have complementary strengths. LMs can be pre-trained on any unstructured text and thus cover a broad scope of information, while KGs are more structured, helping with logical reasoning by connecting randomly generated statements like “People breathe” to logically related ones like “The birthplace of the painter is Italy.

In this research paper, published at NAACL 2021, researchers found that combining both LMs and KGs makes it possible to answer questions more effectively. Existing systems that use LM and KGs tend to be noisy, and the interactions between QA context and KG are not modeled. While this research work offers promising solutions: a) estimating the relevance of nodes in graphs conditioned on the query being asked; b) connecting related elements as joint graph models their relationship.]]></description>
<dc:subject>nlp machine_learning</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:38738c183160/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/dimitreOliveira/bert-as-a-service_TFX">
    <title>dimitreOliveira/bert-as-a-service_TFX: End-to-end pipeline with TFX to train and deploy a BERT model for sentiment analysis.</title>
    <dc:date>2021-06-25T17:48:01+00:00</dc:date>
    <link>https://github.com/dimitreOliveira/bert-as-a-service_TFX</link>
    <dc:creator>amy</dc:creator><description><![CDATA[This repository is designed to demonstrate a simple yet complete machine learning solution that uses a BERT model for text sentiment analysis using a TensorFlow Extended end-to-end pipeline, and making use of some of the best practices from the MLOps domain, it will cover steps from data ingestion until model serving and consuming it either with REST or gRPC requests.]]></description>
<dc:subject>machine_learning google nlp BERT GDEs</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:f2d7be32e445/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:BERT"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:GDEs"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://bit.ly/3bVF0q3">
    <title>AI Powered Misinformation and Manipulation at Scale #GPT-3</title>
    <dc:date>2021-05-28T13:55:54+00:00</dc:date>
    <link>https://bit.ly/3bVF0q3</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Risks of autoregressive language models and the future of prompt engineering]]></description>
<dc:subject>machine_learning NLP</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:090bf9eb1891/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://blog.google/products/search/introducing-mum/">
    <title>MUM: A new AI milestone for understanding information</title>
    <dc:date>2021-05-19T01:02:04+00:00</dc:date>
    <link>https://blog.google/products/search/introducing-mum/</link>
    <dc:creator>amy</dc:creator><dc:subject>nlp google machine_learning search</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:30d488e8d660/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:search"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.wired.com/story/ai-generate-convincing-text-anyone-use-it/?mc_cid=90e7741fa8&amp;mc_eid=eef4b324fd">
    <title>This AI Can Generate Convincing Text—and Anyone Can Use It</title>
    <dc:date>2021-03-30T13:51:52+00:00</dc:date>
    <link>https://www.wired.com/story/ai-generate-convincing-text-anyone-use-it/?mc_cid=90e7741fa8&amp;mc_eid=eef4b324fd</link>
    <dc:creator>amy</dc:creator><description><![CDATA[The makers of Eleuther hope it will be an open source alternative to GPT-3, the well-known language program from OpenAI.]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:fa31dc8273ae/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/pdf/2101.03961.pdf">
    <title>SWITCH TRANSFORMERS: SCALING TO TRILLION PARAMETER MODELS WITH SIMPLE AND EFFICIENT SPARSITY William Fe</title>
    <dc:date>2021-02-16T14:05:52+00:00</dc:date>
    <link>https://arxiv.org/pdf/2101.03961.pdf</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning NLP</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:be065558ceac/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://venturebeat.com/2021/02/09/openai-and-stanford-researchers-call-for-urgent-action-to-address-harms-of-large-language-models-like-gpt-3/">
    <title>OpenAI and Stanford researchers call for urgent action to address harms of large language models like GPT-3</title>
    <dc:date>2021-02-13T08:33:42+00:00</dc:date>
    <link>https://venturebeat.com/2021/02/09/openai-and-stanford-researchers-call-for-urgent-action-to-address-harms-of-large-language-models-like-gpt-3/</link>
    <dc:creator>amy</dc:creator><description><![CDATA[RT @mmitchell_ai: "This further suggests the urgency of using the current time window, during which few actors possess very large language models, to develop appropriate norms and principles for others to follow." Exactly this.]]></description>
<dc:subject>machine_learning ethics NLP</dc:subject>
<dc:source>https://twitter.com/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:62a4aa5e2d7c/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:ethics"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:NLP"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2002.09402">
    <title>[2002.09402] Addressing Some Limitations of Transformers with Feedback Memory</title>
    <dc:date>2021-01-27T16:56:50+00:00</dc:date>
    <link>https://arxiv.org/abs/2002.09402</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Transformers have been successfully applied to sequential, auto-regressive tasks despite being feedforward networks. Unlike recurrent neural networks, Transformers use attention to capture temporal relations while processing input tokens in parallel. While this parallelization makes them computationally efficient, it restricts the model from fully exploiting the sequential nature of the input. The representation at a given layer can only access representations from lower layers, rather than the higher level representations already available. In this work, we propose the Feedback Transformer architecture that exposes all previous representations to all future representations, meaning the lowest representation of the current timestep is formed from the highest-level abstract representation of the past. We demonstrate on a variety of benchmarks in language modeling, machine translation, and reinforcement learning that the increased representation capacity can create small, shallow models with much stronger performance than comparable Transformers.]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:e8a894267b01/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://openai.com/blog/dall-e/">
    <title>DALL·E: Creating Images from Text</title>
    <dc:date>2021-01-06T14:38:20+00:00</dc:date>
    <link>https://openai.com/blog/dall-e/</link>
    <dc:creator>amy</dc:creator><description><![CDATA[DALL·E is a 12-billion parameter version of GPT-3 trained to generate 
images from text descriptions, using a dataset of text–image pairs. 
We’ve found that it has a diverse set of capabilities, including 
creating anthropomorphized versions of animals and objects, combining 
unrelated concepts in plausible ways, rendering text, and applying ]]></description>
<dc:subject>machine_learning nlp transformations to existing images.</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:02b18fdacfc9/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:transformations"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:to"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:existing"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:images."/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://twitter.com/nabla_theta/status/1345130408170541056">
    <title>Leo Gao on Twitter: &quot;Announcing a new dataset: the Pile! A free and publicly available 800GB dataset of diverse English text for language modeling! Download: https://t.co/DDUikP5Igz Paper: https://t.co/9XqVw3y8nY 1/7 https://t.co/glhM7KgBFZ&quot; / Twitter</title>
    <dc:date>2021-01-05T19:07:34+00:00</dc:date>
    <link>https://twitter.com/nabla_theta/status/1345130408170541056</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Announcing a new dataset: the Pile! A free and publicly available 800GB dataset of diverse English text for language modeling!

Download: https://pile.eleuther.ai 
Paper: https://pile.eleuther.ai/paper.pdf]]></description>
<dc:subject>machine_learning nlp big_data</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:795bdca9e810/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:big_data"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/google/patents-public-data/blob/master/models/BERT%20for%20Patents.md">
    <title>patents-public-data/BERT for Patents.md at master · google/patents-public-data</title>
    <dc:date>2020-12-08T20:39:11+00:00</dc:date>
    <link>https://github.com/google/patents-public-data/blob/master/models/BERT%20for%20Patents.md</link>
    <dc:creator>amy</dc:creator><description><![CDATA[The BERT exported here has been trained on >100 million patent documents and was trained on all parts of a patent (abstract, claims, description).

The BERT model exported here comes in two formats:

SavedModel

Checkpoint

The models can also be loaded and saved in another format or just the weights can be saved.]]></description>
<dc:subject>machine_learning nlp google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:d01394b71203/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://cloud.google.com/blog/products/ai-machine-learning/how-ai-improves-patent-analysis">
    <title>How AI improves patent analysis | Google Cloud Blog</title>
    <dc:date>2020-12-08T20:38:02+00:00</dc:date>
    <link>https://cloud.google.com/blog/products/ai-machine-learning/how-ai-improves-patent-analysis</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:aa2d126d2ee1/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/NiuTrans/ABigSurvey">
    <title>GitHub - NiuTrans/ABigSurvey: A collection of 400+ survey papers on Natural Language Processing (NLP) and Machine Learning (ML)</title>
    <dc:date>2020-11-25T20:05:17+00:00</dc:date>
    <link>https://github.com/NiuTrans/ABigSurvey</link>
    <dc:creator>amy</dc:creator><description><![CDATA[⭐️ NLP Survey Papers ⭐️ Getting an initial high-level understanding of different NLP tasks and applications is key. Survey papers help a lot. This repo contains a list of NLP survey papers for getting a bit more exposure to a wide range of NLP tasks. ]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:source>https://apple.com/iphone/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:68867e64418e/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://gradientflow.com/2020nlpsurvey/">
    <title>2020 NLP Survey Report – Gradient Flow</title>
    <dc:date>2020-09-24T19:04:05+00:00</dc:date>
    <link>https://gradientflow.com/2020nlpsurvey/</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:77d2dfeb460b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://bdtechtalks.com/2020/08/17/openai-gpt-3-commercial-ai/amp/">
    <title>The untold story of GPT-3 is the transformation of OpenAI – TechTalks</title>
    <dc:date>2020-09-07T18:12:51+00:00</dc:date>
    <link>https://bdtechtalks.com/2020/08/17/openai-gpt-3-commercial-ai/amp/</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:78ab0c4cc9d4/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/google/making_with_ml/blob/master/semantic_ml/use_sample.js">
    <title>making_with_ml/use_sample.js at master · google/making_with_ml · GitHub</title>
    <dc:date>2020-08-22T00:38:51+00:00</dc:date>
    <link>https://github.com/google/making_with_ml/blob/master/semantic_ml/use_sample.js</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:b434fb4edaa7/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://blog.tensorflow.org/2020/08/introducing-semantic-reactor-explore-nlp-sheets.html?m=1">
    <title>Introducing Semantic Reactor: Explore NLP in Google Sheets — The TensorFlow Blog</title>
    <dc:date>2020-08-22T00:34:33+00:00</dc:date>
    <link>https://blog.tensorflow.org/2020/08/introducing-semantic-reactor-explore-nlp-sheets.html?m=1</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:2513f44fb6e8/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://ai.googleblog.com/2018/05/advances-in-semantic-textual-similarity.html?m=1">
    <title>Google AI Blog: Advances in Semantic Textual Similarity</title>
    <dc:date>2020-08-22T00:34:03+00:00</dc:date>
    <link>https://ai.googleblog.com/2018/05/advances-in-semantic-textual-similarity.html?m=1</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:7560886f1053/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.technologyreview.com/2020/08/14/1006780/ai-gpt-3-fake-blog-reached-top-of-hacker-news">
    <title>A college kid created a fake, AI-generated blog. It reached #1 on Hacker News. | MIT Technology Review</title>
    <dc:date>2020-08-20T17:13:45+00:00</dc:date>
    <link>https://www.technologyreview.com/2020/08/14/1006780/ai-gpt-3-fake-blog-reached-top-of-hacker-news</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:cb1dca3a652e/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.sudowrite.com/">
    <title>Sudowrite</title>
    <dc:date>2020-08-13T13:53:14+00:00</dc:date>
    <link>https://www.sudowrite.com/</link>
    <dc:creator>amy</dc:creator><description><![CDATA[<blockquote>
For the past couple weeks, I've been playing around with Sudowrite, @superamit and friends' GPT3-based text generator for fiction writers. You give it it characters, plot summaries, dialogue or twist endings. 

https://www.sudowrite.com/

1/ https://twitter.com/doctorow/status/1293660117909434368/photo/1
</blockquote>]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:589a34ca6ec8/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://ai.googleblog.com/2020/04/a-scalable-approach-to-reducing-gender.html">
    <title>Google AI Blog: A Scalable Approach to Reducing Gender Bias in Google Translate</title>
    <dc:date>2020-07-30T20:38:23+00:00</dc:date>
    <link>https://ai.googleblog.com/2020/04/a-scalable-approach-to-reducing-gender.html</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp google bias</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:bf1d4747b01b/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:bias"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2002.08909">
    <title>[2002.08909] REALM: Retrieval-Augmented Language Model Pre-Training</title>
    <dc:date>2020-07-21T17:36:02+00:00</dc:date>
    <link>https://arxiv.org/abs/2002.08909</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Language model pre-training has been shown to capture a surprising amount of world knowledge, crucial for NLP tasks such as question answering. However, this knowledge is stored implicitly in the parameters of a neural network, requiring ever-larger networks to cover more facts.
To capture knowledge in a more modular and interpretable way, we augment language model pre-training with a latent knowledge retriever, which allows the model to retrieve and attend over documents from a large corpus such as Wikipedia, used during pre-training, fine-tuning and inference. For the first time, we show how to pre-train such a knowledge retriever in an unsupervised manner, using masked language modeling as the learning signal and backpropagating through a retrieval step that considers millions of documents.
We demonstrate the effectiveness of Retrieval-Augmented Language Model pre-training (REALM) by fine-tuning on the challenging task of Open-domain Question Answering (Open-QA). We compare against state-of-the-art models for both explicit and implicit knowledge storage on three popular Open-QA benchmarks, and find that we outperform all previous methods by a significant margin (4-16% absolute accuracy), while also providing qualitative benefits such as interpretability and modularity.]]></description>
<dc:subject>machine_learning nlp google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:067f002529bc/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://towardsdatascience.com/gpt-3-creative-potential-of-nlp-d5ccae16c1ab">
    <title>GPT-3: Creative Potential of NLP. New ML milestone by OpenAI — in action | by Vlad Alex (Merzmensch) | Jul, 2020 | Towards Data Science</title>
    <dc:date>2020-07-18T16:19:16+00:00</dc:date>
    <link>https://towardsdatascience.com/gpt-3-creative-potential-of-nlp-d5ccae16c1ab</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:identifier>https://pinboard.in/u:amy/b:a6b069bd3217/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://drive.google.com/file/d/19NicWkWcF_-V0UvxUqAzNoRbT-Dp6SPK/view">
    <title>filbert_paper.pdf - Google Drive</title>
    <dc:date>2020-07-16T20:09:49+00:00</dc:date>
    <link>https://drive.google.com/file/d/19NicWkWcF_-V0UvxUqAzNoRbT-Dp6SPK/view</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Measuring and Reducing Gendered Correlations in Pre-trained Models]]></description>
<dc:subject>machine_learning nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:d8ef30839889/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/2004.15011">
    <title>[2004.15011] TLDR: Extreme Summarization of Scientific Documents</title>
    <dc:date>2020-05-05T16:29:08+00:00</dc:date>
    <link>https://arxiv.org/abs/2004.15011</link>
    <dc:creator>amy</dc:creator><description><![CDATA[We introduce TLDR generation for scientific papers, a new automatic summarization task with high source compression, requiring expert background knowledge and complex language understanding. To facilitate research on this task, we introduce SciTLDR, a dataset of 3.9K TLDRs. Furthermore, we introduce a novel annotation protocol for scalably curating additional gold summaries by rewriting peer review comments. We use this protocol to augment our test set, yielding multiple gold TLDRs for evaluation, which is unlike most recent summarization datasets that assume only one valid gold summary. We present a training strategy for adapting pretrained language models that exploits similarities between TLDR generation and the related task of title generation, which outperforms strong extractive and abstractive summarization baselines.]]></description>
<dc:subject>machine_learning nlp summarization</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:1f6353be0418/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:summarization"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://scitldr.apps.allenai.org/">
    <title>SciTLDR</title>
    <dc:date>2020-05-05T16:20:22+00:00</dc:date>
    <link>https://scitldr.apps.allenai.org/</link>
    <dc:creator>amy</dc:creator><description><![CDATA[This demo generates TLDRs for scientific articles, using our best performing model, which works best when given the abstract, introduction, and conclusion of a paper. However, it will still work if you only provide an abstract. Currently, our model is only trained on English-language papers in the Computer Science domain, although we hope future work will expand to more domains/languages!

This is a demo for our paper, "TLDR: Extreme Summarization of Scientific Documents" - https://arxiv.org/abs/2004.15011]]></description>
<dc:subject>machine_learning nlp summarization</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:578c4c052bcb/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:summarization"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1905.07129v3">
    <title>[1905.07129v3] ERNIE: Enhanced Language Representation with Informative Entities</title>
    <dc:date>2020-04-15T21:26:07+00:00</dc:date>
    <link>https://arxiv.org/abs/1905.07129v3</link>
    <dc:creator>amy</dc:creator><description><![CDATA[Neural language representation models such as BERT pre-trained on large-scale corpora can well capture rich semantic patterns from plain text, and be fine-tuned to consistently improve the performance of various NLP tasks. However, the existing pre-trained language models rarely consider incorporating knowledge graphs (KGs), which can provide rich structured knowledge facts for better language understanding. We argue that informative entities in KGs can enhance language representation with external knowledge. In this paper, we utilize both large-scale textual corpora and KGs to train an enhanced language representation model (ERNIE), which can take full advantage of lexical, syntactic, and knowledge information simultaneously. The experimental results have demonstrated that ERNIE achieves significant improvements on various knowledge-driven tasks, and meanwhile is comparable with the state-of-the-art model BERT on other common NLP tasks. The source code of this paper can be obtained from this https URL: https://github.com/thunlp/ERNIE]]></description>
<dc:subject>nlp machine_learning</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:7b7a7c8272d7/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/tensorflow/models/tree/master/official/nlp/xlnet">
    <title>models/official/nlp/xlnet at master · tensorflow/models</title>
    <dc:date>2020-04-15T21:05:37+00:00</dc:date>
    <link>https://github.com/tensorflow/models/tree/master/official/nlp/xlnet</link>
    <dc:creator>amy</dc:creator><description><![CDATA[XLNet is a generalized autoregressive BERT-like pretraining language model that enables learning bidirectional contexts by maximizing the expected likelihood over all permutations of the factorization order. It can learn dependency beyond a fixed length without disrupting temporal coherence by using segment-level recurrence mechanism and relative positional encoding scheme introduced in Transformer-XL. XLNet outperforms BERT on 20 NLP benchmark tasks and achieves state-of-the-art results on 18 tasks including question answering, natural language inference, sentiment analysis, and document ranking.]]></description>
<dc:subject>TensorFlow nlp machine_learning</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:8b07ba058b07/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:TensorFlow"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1906.08237v2">
    <title>[1906.08237v2] XLNet: Generalized Autoregressive Pretraining for Language Understanding</title>
    <dc:date>2020-04-15T20:53:41+00:00</dc:date>
    <link>https://arxiv.org/abs/1906.08237v2</link>
    <dc:creator>amy</dc:creator><description><![CDATA[With the capability of modeling bidirectional contexts, denoising autoencoding based pretraining like BERT achieves better performance than pretraining approaches based on autoregressive language modeling. However, relying on corrupting the input with masks, BERT neglects dependency between the masked positions and suffers from a pretrain-finetune discrepancy. In light of these pros and cons, we propose XLNet, a generalized autoregressive pretraining method that (1) enables learning bidirectional contexts by maximizing the expected likelihood over all permutations of the factorization order and (2) overcomes the limitations of BERT thanks to its autoregressive formulation. Furthermore, XLNet integrates ideas from Transformer-XL, the state-of-the-art autoregressive model, into pretraining. Empirically, under comparable experiment settings, XLNet outperforms BERT on 20 tasks, often by a large margin, including question answering, natural language inference, sentiment analysis, and document ranking.]]></description>
<dc:subject>nlp machine_learning TensorFlow</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:9126ca7eeb61/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:TensorFlow"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.analyticsvidhya.com/blog/2018/03/essentials-of-deep-learning-sequence-to-sequence-modelling-with-attention-part-i/">
    <title>Essentials of Deep Learning – Sequence to Sequence modelling with Attention (using python)</title>
    <dc:date>2020-04-15T20:01:50+00:00</dc:date>
    <link>https://www.analyticsvidhya.com/blog/2018/03/essentials-of-deep-learning-sequence-to-sequence-modelling-with-attention-part-i/</link>
    <dc:creator>amy</dc:creator><dc:subject>nlp machine_learning</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:59c1876f59e5/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.analyticsvidhya.com/blog/2019/06/understanding-transformers-nlp-state-of-the-art-models/">
    <title>Understanding Transformers in NLP: State-of-the-Art Models</title>
    <dc:date>2020-04-15T20:00:18+00:00</dc:date>
    <link>https://www.analyticsvidhya.com/blog/2019/06/understanding-transformers-nlp-state-of-the-art-models/</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:93124d623145/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/zihangdai/xlnet">
    <title>zihangdai/xlnet: XLNet: Generalized Autoregressive Pretraining for Language Understanding</title>
    <dc:date>2020-04-15T19:54:50+00:00</dc:date>
    <link>https://github.com/zihangdai/xlnet</link>
    <dc:creator>amy</dc:creator><description><![CDATA[XLNet: Generalized Autoregressive Pretraining for Language Understanding
]]></description>
<dc:subject>TensorFlow nlp machine_learning</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:c7c65ac6cbc8/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:TensorFlow"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://www.analyticsvidhya.com/blog/2020/03/6-pretrained-models-text-classification/">
    <title>6 Pretrained Models to Master Text Classification</title>
    <dc:date>2020-04-15T19:51:34+00:00</dc:date>
    <link>https://www.analyticsvidhya.com/blog/2020/03/6-pretrained-models-text-classification/</link>
    <dc:creator>amy</dc:creator><dc:subject>machine_learning nlp</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:3abf7cdbd994/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://arxiv.org/abs/1910.10683">
    <title>[1910.10683] Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer</title>
    <dc:date>2020-04-15T17:59:12+00:00</dc:date>
    <link>https://arxiv.org/abs/1910.10683</link>
    <dc:creator>amy</dc:creator><description><![CDATA[T5
Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new "Colossal Clean Crawled Corpus", we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.]]></description>
<dc:subject>nlp machine_learning google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:470e84736203/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
<item rdf:about="https://github.com/google-research/text-to-text-transfer-transformer">
    <title>google-research/text-to-text-transfer-transformer: Code for the paper &quot;Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer&quot;</title>
    <dc:date>2020-03-11T16:49:49+00:00</dc:date>
    <link>https://github.com/google-research/text-to-text-transfer-transformer</link>
    <dc:creator>amy</dc:creator><dc:subject>TensorFlow machine_learning nlp google</dc:subject>
<dc:source>https://pinboard.in/</dc:source>
<dc:identifier>https://pinboard.in/u:amy/b:cf0de93cdb4e/</dc:identifier>
<taxo:topics><rdf:Bag>	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:TensorFlow"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:machine_learning"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:nlp"/>
	<rdf:li rdf:resource="https://pinboard.in/u:amy/t:google"/>
</rdf:Bag></taxo:topics>
</item>
</rdf:RDF>