Safety-Critical Learning Reading Group

Description

Each summer, students and faculty from various DCL focused labs across campus come together to read and discuss papers with a focused theme. This year, the selected topic is provable guarantees for learning-enabled systems. Papers of interest will address topics like

• How should we appropriately define “safety” for such systems (e.g. with semantic understanding)?
• What guarantees of safety can we make, beyond empirical techniques such as domain randomization?
• How should we approach the concept of safety without an explicit model?

Meeting Logistics:

• Location: TSRB 523a
• Time: Wednesdays, 2-3pm
• Dates: Bi-weekly (see below)

If you are interested in attending, please contact Brendan Gould to be added to the mailing list.

Planned Schedule

2026-06-17:

• Paper: RAYEN: Imposition of Hard Convex Constraints on Neural Networks
• Authors: Jesus Tordesills, Victor Klemm, Jonathan P. How, Marco Hutter
• Presenter: Brendan Gould

2025-07-01:

• Paper: Vision-Based Runtime Monitoring under Varying Specifications using Semantic Latent Representations
• Authors: Bardh Hoxha, Oliver Schön, Hideki Okamoto, Lars Lindemann, Georgios Fainekos
• Presenter: Xinpei Ni

2025-07-15:

• Paper: Pick-to-Learn for Systems and Control: Data-driven Synthesis with State-of-the-art Safety Guarantees
• Authors: Dario Paccagnan, Daniel Marks, Marco C. Campi, Simone Garatti
• Presenter: Chih-yuan (Frank) Chiu

2025-07-29:

• Paper: How to Train Your Latent Control Barrier Function: Smooth Safety Filtering Under Hard-to-Model Constraints
• Authors: Kensuke Nakamura, Arun L. Bishop, Steven Man, Aaron M. Johnson, Zachary Manchester, Andrea Bajcsy
• Presenter: Varun Madabushi

2025-08-12:

• Paper: Meta-Adaptive Nonlinear Control: Theory and Algorithms
• Authors: Guanya Shi, Kamyar Azizzadenesheli, Michael O’Connell, Soon-Jo Chung, Yisong Yue
• Presenter: Huaiyuan (Joe) Rao