Uploaded June 2021 | Updated September 2026, 2 weeks ago

---

Content Links:
Break-It-Fix-It: arxiv.org/pdf/2106.06600.pdf
SimGAN: ai.googleblog.com/2021/06/learning-accurate-physics-simulator-via.html
AugLy: ai.facebook.com/blog/augly-a-new-data-augmentation-library-to-help-build-more-robust-ai-models
Scientific LMs for Biomedical KBC: arxiv.org/pdf/2106.09700.pdf
QuIP: arxiv.org/pdf/2106.08190.pdf
Prompting Contrastive Explanations: arxiv.org/pdf/2106.06823.pdf
DocNLI: arxiv.org/pdf/2106.09449.pdf
UnNatural Language Inference: arxiv.org/pdf/2101.00010.pdf
Multi-head or Single Head? arxiv.org/pdf/2106.09650.pdf
GNNs for NLP: arxiv.org/abs/2106.06090
Understandin VQ-VAE: https://ml.berkeley.edu/blog/posts/vq-vae/
BEIT: arxiv.org/pdf/2106.08254.pdf
Long-Short Temporal Contrastive Learning: arxiv.org/pdf/2106.09212.pdf
Watching Too Much TV: arxiv.org/pdf/2106.08513.pdf
DL Through the Lens of Example Difficulty: arxiv.org/pdf/2106.09647.pdf
Pre-trained Models: arxiv.org/pdf/2106.07139.pdf
Counterfactual Explanations in ML: arxiv.org/pdf/2106.07756v1.pdf
Reverse Engineering of Generative Models: arxiv.org/pdf/2106.07873.pdf
Thinking like Transformers: arxiv.org/abs/2106.06981
HuggingFace Course: huggingface.co/course/chapter1
HF Jax Backend LM Training: github.com/huggingface/transformers/tree/master/examples/flax/language-modeling
Style Transfer on Model Hub: huggingface.co/prithivida/informal_to_formal_styletransfer
KCE - Video Classification with Transformers: keras.io/examples/vision/video_transformers
Geometric Deep Learning: youtube.com/watch?v=w6Pw4MOzMuo
Learning from Limited and Imperfect Data: youtube.com/watch?v=7CmALEnXee8&list=PLFFnmJXKA-xz6DRF3ZaftkRZS8bOaYzHR&index=6
Transformers v4.7.0: twitter.com/huggingface/status/1405561474949345287
Model Hub Dataset: kaggle.com/crazydiv/huggingface-modelhub
FEVEROUS: paperswithcode.com/dataset/feverous
RobustBench v0.2: twitter.com/fra__31/status/1405588311306125326
LARC: paperswithcode.com/dataset/larc
Learning to Predict Visual Attributes in the Wild: arxiv.org/pdf/2106.09707.pdf
PapersWithCode Newsletter: paperswithcode.com/newsletter/11

Chapters
0:00 Intro
11:02 Break-It-Fix-It
13:24 SimGAN
14:56 AugLy
15:47 Scientific LMs for Biomedical KBC
17:13 QuIP
18:52 Prompting Contrastive Explanations
19:47 DocNLI
20:26 UnNatural Language Inference
21:18 Multi-head or Single Head?
22:22 GNNs for NLP
22:54 BEIT
24:35 Long-Short Temporal Contrastive Learning and Watching Too Much TV
26:10 DL Through the Lens of Example Difficulty
26:52 Pre-trained Models
28:37 Counterfactual Explanations in ML
29:37 Reverse Engineering of Generative Models
30:20 Thinking like Transformers
30:50 HuggingFace Course
31:28 HF Jax Backend LM Training
32:32 Style Transfer on Model Hub
33:10 KCE - Video Classification with Transformers
33:33 Recommended Talks
34:17 Transformers v4.7.0
34:37 Model Hub Dataset
35:23 New Datasets
36:33 PapersWithCode Newsletter

Thanks for watching! Please Subscribe!