Workshop Schedule

The workshop will be held in Baltimore, MD, United States, co-located with ICML 2022.

Workshop date: July 22 (Friday), full-day (8:45 am ET - 6 pm ET).

Time zone: US Eastern Time (UTC -5)

Location: Baltimore Convention Center, Room 308

8:45 am - 9 am: Opening remarks

Morning Session 1:

9:30 am - 9:45 am: Contributed Talk 1: "Backward Reachability for Neural Feedback Loops"

9:45 am - 10 am: Contributed Talk 2: "Characterizing Neural Network Verification for Systems with NN4SysBench"

10 am - 10:30 am: Break

Morning Session 2:

10:30 am - 11 am: Invited Talk 2: Dr. Anton Dahbura (Johns Hopkins University) - Undeterminism and the AI Uncertainty Principle

11 am - 11:30 am: Invited Talk 3: Dr. M. Pawan Kumar (DeepMind) - Neural Networks for Neural Network Verification

11:30 am - 12:00 pm: Morning Poster Session: Some papers will be presented on Zoom; find zoom links here

12:00 pm - 1:30 pm: Lunch Break

Afternoon Session 1:

1:30 pm - 2 pm: Invited Talk 4: Prof. Suman Jana (Columbia University) - Efficient Neural Network Verification using Branch and Bound

2 pm - 2:30 pm: Invited Talk 5: Prof. Somayeh Sojoudi (UC Berkeley) - Computational Methods for Non-convex Machine Learning Problems

2:30 pm - 2:45 pm: Contributed Talk 3: "IBP Regularization for Verified Adversarial Robustness via Branch-and-Bound"

2:45 pm - 3 pm: Contributed Talk 4: "Improved Certified Defenses against Data Poisoning with (Deterministic) Finite Aggregation"

3 pm - 3:30 pm: Break

Afternoon Session 2:

3:30 pm - 4 pm: Invited Talk 6: Prof. Changliu Liu (CMU) - Applications of Neural Verification on Robotics

4 pm - 4:15 pm: Contributed Talk 5: "Towards Optimal Randomized Smoothing: A Semi-Infinite Linear Programming Approach"

4:15 pm - 4:30 pm: Contributed Talk 6: "Don't Lie to Me! Robust and Efficient Explainability with Verified Perturbation Analysis"

4:30 pm - 5 pm: Afternoon Poster Session: Some papers will be presented on Zoom; find zoom links here

Panel:

5 pm - 5:45 pm: Panel (Prof. Gagandeep Singh, Dr. Anton Dahbura, Dr. M. Pawan Kumar and Prof. Changliu Liu)

5:45 pm - 6 pm: Closing Remarks