Frontiers of Deep Learning

Program

Foundations of Deep Learning

Location

Calvin Lab auditorium

Date

Monday, July 15 – Thursday, July 18, 2019

Back to calendar

8:30 – 8:50 a.m.

Coffee and Check-In

8:50 – 9 a.m.

Opening Remarks

9 – 9:45 a.m.

Benign Overfitting in Linear Prediction

Peter Bartlett (Simons Institute, UC Berkeley)

9:45 – 10:30 a.m.

Practical Model-based Algorithms for Reinforcement Learning and Imitation Learning, with Theoretical Analyses

Tengyu Ma (Facebook AI Research)

10:30 – 11 a.m.

Break

11 – 11:45 a.m.

Analyzing Optimization and Generalization in Deep Learning via Trajectories of Gradient Descent

Nadav Cohen (Institute for Advanced Study)

11:45 a.m. – 12:30 p.m.

From Classical Statistics to Modern Machine Learning

Misha Belkin (UCSD)

12:30 – 2:30 p.m.

Lunch

2:30 – 3:15 p.m.

Learning and Generalization in Over-parametrized Neural Networks, Going Beyond Kernels

Yuanzhi Li (Stanford University)

3:15 – 4 p.m.

Size-free Generalization Bounds for Convolutional Neural Networks

Hanie Sedghi (Google Brain)

4 – 5 p.m.

Reception

8:30 – 9 a.m.

Coffee and Check-In

9 – 9:45 a.m.

On the Foundations of Deep Learning: SGD, Overparametrization, and Generalization

Jason Lee (University of Southern California)

9:45 – 10:30 a.m.

A Primal-dual Analysis of Margin Maximization by Steepest Descent Methods

Matus Telgarsky (New York University)

10:30 – 11 a.m.

Break

11 – 11:45 a.m.

Explaining Landscape Connectivity of Low-cost Solutions for Multilayer Nets

Rong Ge (Duke University)

11:45 a.m. – 12:30 p.m.

Training on the Test Set and Other Heresies

Benjamin Recht (UC Berkeley)

12:30 – 2:30 p.m.

Lunch

2:30 – 3:15 p.m.

Computation in Very Wide Neural Networks

Yasaman Bahri (Google Brain)

3:15 – 4 p.m.

Mad Max: Affine Spline Insights into Deep Learning

Richard Baraniuk (Rice University)

4 – 4:30 p.m.

Break

4:30 – 5:15 p.m.

Splitting Gradient Descent for Incremental Learning of Neural Architectures

Qiang Liu (University of Texas at Austin)

8:30 – 9 a.m.

Coffee and Check-In

9 – 9:45 a.m.

Provable Robustness Beyond Bound Propagation

Zico Kolter (Carnegie Mellon University)

9:45 – 10:30 a.m.

A New Perspective on Adversarial Perturbations

Aleksander Mądry (Massachusetts Institute of Technology)

10:30 – 11 a.m.

Break

11 – 11:45 a.m.

Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers

Jerry Li (Microsoft Research)

11:45 a.m. – 12:30 p.m.

Interpreting Deep Neural Networks (DNNs)

Bin Yu (UC Berkeley)

12:30 – 2:30 p.m.

Lunch

2:30 – 3:15 p.m.

Interpretability - now what?

Been Kim (Google Brain)

3:15 – 4 p.m.

Lessons Learned from Evaluating the Robustness of Defenses to Adversarial Examples

Nicholas Carlini (Google Brain)

4 – 4:30 p.m.

Break

4:30 – 5:30 p.m.

Plenary Discussion: Frontiers of Deep Learning

8:30 – 9 a.m.

Coffee and Check-In

9 – 9:45 a.m.

Robust Deep Learning Under Distribution Shift

Zack Lipton (Carnegie Mellon University)

9:45 – 10:30 a.m.

Using Deep Learning for Perception in Autonomous Systems: A Perspective from Control Theory

Claire Tomlin (UC Berkeley)

10:30 – 11 a.m.

Break

11 – 11:45 a.m.

Kernel and Deep Regimes in Overparameterized Learning

Suriya Gunasekar (Toyota Technology Institute, Chicago)

11:45 a.m. – 12:30 p.m.

Towards Understanding Transfer Learning with Applications to Medical Imaging

Maithra Raghu (Cornell University and Google Inc.)

12:30 – 2:30 p.m.

Lunch

2:30 – 3:15 p.m.

Meta-learning of Optimizers and Update Rules

Jascha Sohl-Dickstein (Google Brain)

3:15 – 4 p.m.

Studying Generalization in Deep Learning via PAC-Bayes

Gintare Karolina Dziugaite (Element AI)

4 – 4:15 p.m.

Break

4:15 – 4:45 p.m.

Sample-complexity of Estimating Convolutional and Recurrent Neural Networks

Aarti Singh (Carnegie Mellon University)