Schedule

Note: the poster session and panel discussion will take place at these specific time slots (Pacific time zone). All other talks will be pre-recorded.

Session 1

• 7:30am - 7:40am Opening remarks by Sharon Yixuan Li

• 7:40am - 8:10 Invited talk by Matthias Hein Towards reliable and robust machine learning

• 8:10am - 9:00am Spotlight Session

• 9:00am - 10:00am Poster Session

Coffee Break: 10:00-10:30am

Session 2

• 10:30am - 11:00 Invited talk by Finale Doshi-Velez Uncertainty in Deep Learning: How to be Bayesian?

• 11:00am - 11:30 Invited talk by Percy Liang Tradeoffs between Robustness and Accuracy

• 11:30am - 12:30pm Panel Discussion

Lunch: 12:30-1:30pm

Session 3

• 1:30pm - 2:00pm Invited talk by Raquel Urtasun Uncertainty and Robustness for Self-driving

• 2:00pm - 2:10pm Contributed talk #1 Confidence-Calibrated Adversarial Training: Generalizing to Unseen Attacks

• 2:10pm - 2:20pm Contributed talk #2 Improving robustness against common corruptions by covariate shift adaptation

• 2:20pm - 2:30pm Contributed talk #3 A Unified View of Label Shift Estimation

• 2:30pm - 3:00pm Invited talk by Justin Gilmer Why Adversarial Examples Feel Like Bugs

Coffee Break: 3:00-3:30pm

Session 4

• 3:30pm - 3:40pm Contributed talk #4 A Benchmark of Medical Out of Distribution Detection

• 3:40pm - 3:50pm Contributed talk #5 Neural Ensemble Search for Performant and Calibrated Predictions

• 3:50pm - 4:00pm Contributed talk #6 Bayesian model averaging is suboptimal for generalization under model misspecification