RLC WORKSHOP

Workshop on Programmatic Reinforcement Learning

Aug 5, 2025

Edmonton, Canada

This workshop explores using programmatic representations (e.g., code, symbolic programs, rules) to enhance agent learning and address key challenges in reinforcement learning (RL). By leveraging structured representations, we aim to improve interpretability, generalization, efficiency, and safety in deep RL, moving beyond the limitations of “black box” deep learning models. The workshop brings together researchers in RL and program synthesis/code generation to discuss using programs as policies (e.g., LEAPS, Code as Policies, HPRL, RoboTool, Carvalho et al. 2024), reward functions (e.g., Eureka, Language2Reward, Text2Reward), skill libraries (e.g., Voyager), task generators (e.g., GenSim), or environment models (e.g., WorldCoder, Code World Models). This paradigm bridging RL and programmatic representations enables human-understandable reasoning, reduces reliance on massive data-driven models, and promotes modularity, fostering progress toward verifiable and robust agents across virtual and real-world applications.

Tentative Schedule

Time	Event
9:00--9:10	Opening remarks
9:10--9:40	Invited talk 1
9:40--10:10	Invited talk 2
10:10--10:45	Oral talks
10:45--11:00	Coffee break
11:00--11:30	Invited talk 3
11:30--12:00	Invited talk 4
12:00--13:00	Lunch
13:00--14:00	Poster session 1
14:00--14:30	Invited talk 5
14:30--15:00	Invited talk 6
15:00--16:00	Poster session 2
16:00--16:15	Coffee break
16:15--17:00	Panel discussion

All times are in Pacific Time (PT).

Speakers

Sheila McIlraith (tentative)

University of Toronto

University of Alberta

Amy Zhang

UT Austin

Jacky Liang (tentative)

Google DeepMind

Organizers

Shao-Hua Sun

National Taiwan University

Levi Lelis

University of Alberta

Call For Papers

Open Review

We invite the submission of research papers and position papers on the topic of programmatic representations for reinforcement learning and sequential decision-making. This workshop aims to bring together researchers from reinforcement learning, imitation learning, planning, search, and optimal control with experts in program synthesis and code generation to explore the use of programmatic structures to enhance agent learning.

Topics of Interest

Topics of interest include, but are not limited to:

Programs as Policies: Representing decision-making logic through programmatic policies in Python or domain-specific languages to enhance generalization and interpretability.
Programs as Reward Functions: Synthesizing programs encoding reward functions for agent learning to alleviate the efforts of designing hand-crafted reward functions.
Programs as Skill Libraries: Representing acquired skills as programs, allowing for reusing and composing skills.
Programmatically Generating Tasks: Synthesizing programs that describe various tasks for agent and robot learning to increase the generalizability of learned policies.
Programs as Environment Models: Inferring executable codes to simulate environment dynamics and using them to get imagined data or plan as in model-based RL.

Submission Format: We accept submissions up to 8 pages in RLC format.

Important Dates

Submission Deadline: May 30, 2025, AoE
Author Notification: June 15, 2025, AoE
Camera-Ready Deadline: July 25, 2025, AoE
Workshop Date: August 5, 2025

Submission Process

All submissions will be managed through OpenReview with a double-blind review process. Accepted papers will be presented during poster sessions, with exceptional submissions selected for oral presentations.

Please submit your paper to the Open Review page.

This website is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License and is based on the Leela Interp project. That means you're free to borrow the source code of this website with attribution.