Program of the day

LOCATION: Campus Center at hotel UMass: https://maps.app.goo.gl/LY9xwmcUQHAcFu9a7 , orals in Room 162 and posters in Room 165.

• 9:00 - 9:15: Welcome remarks and Introduction

• 9:15 - 9:45: Keynote Stephanie Milani (Carnegie Mellon University) - Intuitive, Interpretable, and Aligned: Incorporating Human Factors in Reinforcement Learning

• 9:45 - 10:15: Keynote Sonali Parbhoo (Imperial College London) - Interpretable RL in Action Decision Making in Healthcare.

• 10:15 - 10:35: Lightning orals #1

• 10:35 - 11:00 Coffee break (Conference wide)

• 11:00 - 11:20: Lightning orals #2

• 11:20 - 12:00: Poster session

• 12:00 - 13:00: Lunch break (Location TBD)

• 13:00 - 13:30: Keynote Osbert Bastani (University of Pennsylvania) - Neursymbolic Reinforcement Learning

• 13:30 - 14:00: Lightning orals #3

• 14:00 - 14:30: Poster session

• 14:30 - 15:00: Coffee break (Conference wide)

• 15:00 - 15:15: Poster session

• 15:15 - 15:45: Keynote Tianpei Yang (University of Alberta, online) - Interpretable Reinforcement Learning through Inductive Logic Programming

• 15:45 - 16:00: Closing statement & feedback

Orals

Authors will present their papers in a short 4 min presentation with no audience questions to encourage chats around their posters.

Lightning orals #1 (10:15 - 10:35):

• Zahra Bashir - Assessing the Interpretability of Programmatic Policies using Large Language Models

• Jannis Blüml - OCAtari: Object-Centric Atari 2600 Reinforcement Learning Environments

• Quentin Delfosse - Interpretable Concept Bottlenecks to Align Reinforcement Learning Agent

• Ronald Parr - Transition Noise Facilitates Interpretability

Lightning orals #2 (11:00 - 11:20):

• David DeFazio - Learning Quadruped Locomotion Policies Using Logical Rules

• Paul Festor - Evaluating the impact of explainable RL on physician decision-making in high-fidelity simulations: insights from eye-tracking metrics

• Benjamin Fuhrer - Gradient Boosting Reinforcement Learning

• Zhuorui Ye - Concept-Based Interpretable Reinforcement Learning with Limited to No Human Labels

Lightning orals #3 (13:30 - 14:00):

• Bjarne Gregori - HackAtari: Atari Learning Environments for Robust and Continual Reinforcement Learning

• Ann Huang - Learning Dynamics and the Geometry of Neural Dynamics in Recurrent Neural Controllers

• Sonja Johnson-Yu - Understanding biological active sensing behaviors by interpreting learned artificial agent policies

• Yoann Poupart - Contrastive Sparse Autoencoders for Interpreting Planning of Chess-Playing Agents

• Maxime Wabartha - Piecewise Linear Parametrization of Policies: Towards Interpretable Deep Reinforcement Learning

• Liang Zhang - Deep Reinforcement Learning with Vector Quantized Encoding

Speakers