Program
Schedule
Below is a tentative schedule for the workshop. Invited speakers will give 25 minute talks what safety is and different approaches to it in robotics. These talks will aim to provide broad perspectives on how to define safety, what are the limits on current guarantees, and what challenges should the community focus on.
Date: Sunday Jan 23rd
Location: Building 101, Room 00036
9:00 - 9:15
9:15 - 9:45
9:45 - 10:00
10:00 - 10:30
10:30 - 11:00
11:00 - 11:15
11:15 - 11:45
11:45 - 12:00
12:00 - 12:15
12:15 - 13:30
13:30 - 14:00
14:00 - 14:30
14:30 - 15:30
15:30 - 16:00
16:00 - 16:15
16:15 - 16:45
16:45 - 17:00
17:00 - 17:30+
Welcome and opening remarks
George Pappas -- Title TBA
Spotlight talks session 1
Poster session 1 and coffee break (all posters in 101-0019)
Morteza Lahijanain -- Safety in Face of Complexity: an Algorithmic Approach
Contributed talk 1: Safe Learning and Control using Meta-Learning (James Harrison)
Dorsa Sadigh -- Losing my Rigidity: A journey starting from Formal Methods
Contributed talk 2: Learning-based Model Predictive Control for Safe Reinforcement Learning (Tortsen Koller)
Contributed talk 3: Learning Parametric Constraints in High Dimensions from Demonstrations (Glen Chou)
Lunch break
Melanie Zeilinger -- Imposing Safety Constraints in Learning-based Control
Russ Tedrake -- Verification and validation for robust manipulation
Breakout session and coffee break
Jack Zhu & Teo Tomic -- Fully Autonomous Flying Robots in the Wild: Progress and Challenges
Contributed talk 4: A Scalable Risk-based Framework for Rigorous Autonomous Vehicle Evaluation (Aman Sinha and Matthew O'Kelly)
Christian Pek -- Towards a Future of Self-Verifying Robots
Spotlight talks session 2
Brief closing remarks and poster session 2 (all posters in robotics hall 2)
Session 1: Spotlight Talks & Poster Session
- Safe Learning and Control using Meta-Learning
- Learning Parametric Constraints in High Dimensions from Demonstrations
- Strategy Synthesis in POMDPs via Game-Based Abstractions
- Sensor-based complex mission planning for autonomous robots in unknown environments
- Keep soft robots soft - a data-driven based trade off between feed-forward and feedback control
Session 2: Spotlight Talks & Poster Session
- Data-driven construction of robust motion primitives for non-holonomic vehicles
- Towards Provably Correct Driver Assistance Systems through Stochastic Cognitive Modeling
- The Adaptive Stress Testing Formulation
- A Scalable Risk-based Framework for Rigorous Autonomous Vehicle Evaluation
- Safety with uncertainty specifications
- Learning-based model predictive control for safe reinforcement learning
Accepted Papers
Accepted papers will be presented during interactive poster sessions. Each paper will present a short pitch talk before the poster session. Two papers will be presented in longer form as part of the contributed talks.