Uploaded March 2021 | Updated September 2026, 3 weeks ago

---

Thanks for watching! Please Subscribe!

Content Links:
Revisiting ResNets: arxiv.org/pdf/2103.07579.pdf
Is it Enough to Optimize CNN Architectures on imageNet? arxiv.org/pdf/2103.09108.pdf
Learning to Resize Images for Computer Vision Tasks: arxiv.org/pdf/2103.09950.pdf
Large-Scale Zero-Shot Learning: arxiv.org/pdf/2103.09669.pdf
Training GANs with Stronger Augmentations via Contrastive Discriminator: arxiv.org/pdf/2103.09742.pdf
Using Latent Space Regression to Analyze and Leverage Compositionality in GANs: arxiv.org/pdf/2103.10426.pdf
Greedy Hierarchical Variational Autoencoders for Large-Scale Video Prediction: sites.google.com/view/ghvae
Unsupervised Image Transformation Learning via Generative Adversarial Networks: arxiv.org/pdf/2103.07751.pdf
How Many Data Points is a Prompt Worth? arxiv.org/pdf/2103.08493.pdf
GPT Understands, Too: arxiv.org/pdf/2103.10385.pdf
Few-Shot Text Classification: arxiv.org/pdf/2103.07552.pdf
All NLP Tasks are Generation Tasks: arxiv.org/pdf/2103.10360.pdf
TimeSformer: ai.facebook.com/blog/timesformer-a-new-architecture-for-video-understanding
PapersWithCode Newsletter: paperswithcode.com/newsletter/6

Chapters
0:00 Preview
13:37 Revisiting ResNets
20:40 Is it Enough to Optimize CNN Architectures on ImageNet?
24:27 Learning to Resize
28:27 Large-Scale Zero-Shot Learning
30:51 GANs with Contrastive Discriminator
34:28 Latent Space Regressor
36:38 Greedy Hierarchical Variational Autoencoder
38:43 Unsupervised Image Translation
42:14 How Many Data points is a Prompt Worth?
45:37 GPT Understands, Too
47:13 Few-Shot Text Classification with Triplet Networks, Data Augmentation, and Curriculum Learning
48:31 All NLP Tasks are Generation Tasks
49:48 TimeSformer
51:17 PapersWithCode Newsletter