Thank you, MILA, for articulating why most NLP researchers have been so frustrated by the wave of GANs-for-text papers: https://t.co/fd0EMGQjMZ
β Sam Bowman (@sleepinyourhat) November 7, 2018
Thank you, MILA, for articulating why most NLP researchers have been so frustrated by the wave of GANs-for-text papers: https://t.co/fd0EMGQjMZ
β Sam Bowman (@sleepinyourhat) November 7, 2018
In case you missed it, our first techblog post @CuraiHQ describes applications, history, and research trends in #NLP for #Healthcare https://t.co/tiXcztwcYc
β Xavierππ€π (@xamat) November 6, 2018
BlackBox NLP 2018 slides: https://t.co/DL0smjHchA
β (((Ω()(Ω() 'yoav)))) (@yoavgo) November 5, 2018
Here is an op-for-op @PyTorch re-implementation of @GoogleAI's BERT model by @sanhestpasmoi, @timrault and I.
β Thomas Wolf (@Thom_Wolf) November 5, 2018
We made a script to load Google's pre-trained models and it performs about the same as the TF implementation in our tests (see the readme).
Enjoy!https://t.co/dChmNPGPKO
Here are the slides of my talk on Transfer learning with language models at the Belgium NLP meetup last week. I tried to distill our current understanding of what LMs capture.https://t.co/01l1ysXOQl pic.twitter.com/ME2Vt7Lm7H
β Sebastian Ruder (@seb_ruder) November 5, 2018
We have released @TensorFlow code+models for BERT, a brand new pre-training technique which is now state-of-the-art on a wide array of natural language tasks. It can also be used on many new tasks with minimal changes and quick training! https://t.co/rLR6U7uiPj
β Google AI (@GoogleAI) November 2, 2018
"You may not need attention" by Press and Smith with PyTorch code at https://t.co/sqBZ1CFq2u https://t.co/2oxk8OORPt
β PyTorch (@PyTorch) November 1, 2018
Yandex Data School: Course in Natural Language Processing
β ML Review (@ml_review) October 31, 2018
Github Course by Russian Search Giant
* lectures
* seminars
* home assignmentshttps://t.co/Y4QheAvTuV pic.twitter.com/2Pra0a51d8
Code and pretrained weights for BERT are out now.
β Sebastian Ruder (@seb_ruder) October 31, 2018
Includes scripts to reproduce results. BERT-Base can be fine-tuned on a standard GPU; for BERT-Large, a Cloud TPU is required (as max batch size for 12-16 GB is too small).https://t.co/CWv8GMZiX5
A Keras implementation of BERT -- a new transformer architecture with strong performance across a range of language tasks. https://t.co/OznxM3h51Y
β FranΓ§ois Chollet (@fchollet) October 30, 2018
This text classification example from TensorFlow.js (train a model locally in Keras, deploy it in the browser) is excellent: https://t.co/hrrVpRgxEF
β Josh Gordon (@random_forests) October 29, 2018
Check out the #conll2018 best papers:
β Sebastian Ruder (@seb_ruder) October 29, 2018
- Artetxe et al. reveal that word embeddings capture more information than typically assumed. https://t.co/vvN4sVtXYR
- Barrett et al. show that human attention is a good inductive bias for attention. https://t.co/dvNLj9HGjE https://t.co/QHRZdqRVnX