GenAI: Generation and Generalization

ICLR'20 Social on Generative ML


The theme is Deep Generative Models (DGM) and their applications- generation (by NN), data augmentation (for NN), and generalization (of NN).
We would like to focus the talks on cutting-edge insights, our goal is to enable experience-sharing between intermingled areas and to foster cross-disciplinary work.

Topics of interest: GAN/VAE (latest theory, surveys),  Transfer Learning, Data manipulation (Augmentation, Enhancement),  etc.
Schedule: 28th April (Tuesday), 1600-1800 GMT (i.e. 6-8 pm CEST or 12-2 pm EDT)

Speakers

Google Brain

MIT (CSAIL)

Google Research

Cornell/
ASAPP

Schedule 

(All times in GMT/UTC +0)

16:00 - 16:05 Introduction (Prabhu and Mehdi)

16:05 - 16:30 Keynote | Generalization in Generative Models (Oriol Vinyals- DeepMind)

16:30 - 16:45 Generative models as data visualization (Phillip Isola- MIT)

16:45 - 17:00 Feature Normalization & Data Augmentation (Kilian Weinberger- Cornell)

17:00 - 17:15 Better Generalization for Gen.ML (Puneet Dokania- Oxford)

17:15 - 17:30 Few-Shot Learning: A Benchmark Problem for Gen.ML? (Hugo Larochelle- Google Brain)

17:30 - 17:45 Ethical Considerations of Generative AI (Emily Denton - Google Research)

17:45 - 18:30 ----- Q&A----- Zoom: breakout-rooms

Max Planck Institute for Intelligent Systems (MPI-IS)
Tübingen, Germany

Google Brain
Berlin, Germany

Thanks to our Social Chair- Adam White, and a big shoutout for all the ICLR Volunteers!