Schedule
Schedule*
07:55 - 08:00 Welcome
08:00 - 08:40 Invited Talk 1: Isabelle Guyon
08:40 - 09:00 Q&A session (Invited Talk 1)
09:00 - 09:40 Invited Talk 2: Song Han
09:40 - 10:00 Q&A session (Invited Talk 2)
10:00 - 10:15 Contributed Talk 1: Variational Depth Search in ResNets
Javier Antoran, James Urquhart Allingham and José Miguel Hernández-Lobato
One-shot neural architecture search allows joint learning of weights and network architecture, reducing computational cost. We limit our search space to the depth of residual networks and formulate an analytically tractable variational objective that allows for obtaining an unbiased approximate posterior over depths in one-shot. We propose a heuristic to prune our networks based on this distribution. We compare our proposed method against manual search over network depths on the MNIST, Fashion-MNIST, SVHN datasets. We find that pruned networks do not incur a loss in predictive performance, obtaining accuracies competitive with unpruned networks. Marginalising over depth allows us to obtain better-calibrated test-time uncertainty estimates than regular networks, in a single forward pass.
10:15 - 10:30 Q&A session (Contributed Talk 1)
10:30 - 11:45 Virtual Poster Session (all papers)
11:45 - 12:25 Invited Talk 3: Ameet Talwalkar
12:25 - 12:45 Q&A session (Invited Talk 3)
12:45 - 13:25 Invited Talk 4: Quoc V. Le. AutoML at Google and Future Directions
13:25 - 13:45 Q&A session (Invited Talk 4)
13:45 - 14:00 Contributed Talk 2: A Closer Look at Reinforcement Learning for Neural Network Architecture Search
James A. Preiss, Eugen Hotaj and Hanna Mazzawi
We explore the design decisions involved in reducing neural network architecture search (NAS) to a reinforcement learning (RL) problem. We compare several reductions on the NAS-Bench-101 dataset, while holding the RL algorithm and search space constant. Based on our findings, we discuss how NAS differs from typical RL settings, and suggest guidelines for applying RL to NAS problems.
14:00 - 14:15 Q&A session (Contributed Talk 2)
14:15 - 15:00 Panel Discussion
- Quoc V. Le (Google Brain)
- Song Han (MIT)
- Ameet Talwalkar (Carnegie Mellon University & Determined AI)
- Isabelle Guyon (University of Paris-Sud, INRIA, ChaLearn & ClopiNet)
Please add questions for the panel here.