Google: BERT now used on almost every English query https://t.co/dO4AaFkNdt
— Bojan Tunguz (@tunguz) October 17, 2020
Google: BERT now used on almost every English query https://t.co/dO4AaFkNdt
— Bojan Tunguz (@tunguz) October 17, 2020
While language models are capable of generating coherent texts, they can internalise social biases present in the training corpus. Bias extends to attributes (e.g., sentiment) of the generated text, though this can be reduced: https://t.co/B1ly3O2uwD, in Findings of #emnlp2020 pic.twitter.com/4N3HDslLDn
— DeepMind (@DeepMind) October 16, 2020
*Vokenization*: a visually-supervised language model attempt in our #emnlp2020 paper: https://t.co/r9MZNniAhn (w. @mohitban47)
— Hao Tan (@HaoTan5) October 15, 2020
To improve language pre-training, we extrapolate multimodal alignments to lang-only data by contextually mapping tokens to related images ("vokens") 1/4 pic.twitter.com/wuXt1K58BH
Happy to share our latest paper: "Self-training Improves Pretraining for Natural Language Understanding"
— Alexis Conneau (@alex_conneau) October 15, 2020
We show that self-training is complementary to strong unsupervised pretraining (RoBERTa) on a variety of tasks.
Paper: https://t.co/Fi1N9UKao7
Code: https://t.co/SsPSENYw5L pic.twitter.com/n4IUsYfVGF
Ceyda Cinarel (@ceyda_cinarel) shared a neat demo of a custom inference widget for token classification models 🏷, built on top of the:
— Julien Chaumond (@julien_c) October 9, 2020
- @huggingface Inference API
- Displacy from @spacy_io
- @streamlit
🔥 Demo: https://t.co/Rlk5gb9rO0
🔥 Blogpost: https://t.co/rKwlrFeMj2
🗂 We published a new collection of BERT models!
— TensorFlow (@TensorFlow) October 8, 2020
Models are trained on different datasets and tasks including: PubMed, MNLI, and SQuAD to boost downstream tasks performance.
See more → https://t.co/hmxBBBCuPj
Try the tutorial → https://t.co/eMd3YWzUaE pic.twitter.com/ND2BbWBTIf
Megatron-CTRL: Controllable Story Generation using External Knowledge from @NVIDIAAI at #emnlp #gtc2020 talk Thursday at 11PM PDT. @pengxu1026 Mostafa Patwary @MohammadShoeybi @TheRealRPuri @pascalefung @ctnzr
— Prof. Anima Anandkumar (@AnimaAnandkumar) October 7, 2020
Blog: https://t.co/5816nwruql
Paper: https://t.co/RExkIEEdvp
The ultimate guide to encoder-decoder models!
— Hugging Face (@huggingface) October 5, 2020
Today, we're releasing part one explaining how they work and why they have become indispensable for NLG tasks such as summarization and translation.
> https://t.co/MbIRQMVtzU
Subscribe for the full series: https://t.co/O2BN6nth3I pic.twitter.com/LxOC8qekbg
New hot pre-print 🔥Interpreting Graph Neural Networks for NLP With Differentiable Edge Masking🔥 https://t.co/1eRIMPzttX
— Nicola De Cao (@nicola_decao) October 2, 2020
We show you can learn to remove most of the edges in GNNs such that the remaning ones are interpretable!
with @michael_sejr @iatitov pic.twitter.com/0xl3U8bxiB
We're excited to announce the🤗Transformers release of the Retrieval-Augmented Generation model in collaboration with @facebookai!
— Hugging Face (@huggingface) September 28, 2020
Paper: https://t.co/KgiUdQ8Gzg
Demo: https://t.co/RigCiHuqTK
🤗Doc: https://t.co/o2bUBmzLvJ
Blog post: https://t.co/J18sYTa6Da pic.twitter.com/NBjy4tEjSz
Transformer’s attention mechanism can be linked to other cool ideas in AI
— hardmaru (@hardmaru) September 26, 2020
- Indirect Encoding in Neuroevolutionhttps://t.co/G740mhjBv4
- Hopfield Networks with continuous stateshttps://t.co/FL8PimjVo9
- Graph Neural Networks with multi-head attentionhttps://t.co/PACMnKT50F
Today we describe a #NaturalLanguageProcessing model that achieves near BERT-level performance on text classification tasks, while using orders of magnitude fewer model parameters. Learn all about it below: https://t.co/94GZU4GOt3
— Google AI (@GoogleAI) September 21, 2020