RLC 2025 Accepted Workshops

Explores programmatic representations — symbolic programs, code-based policies, and abstractions — to address interpretability, generalizability, efficiency, and safety in reinforcement learning.

Celebrates the 10th anniversary of the Deep Q-Network Atari Nature paper, discussing the intertwined history and future research problems of RL in video games.

Investigates how encoding prior knowledge via inductive biases can boost performance, sample efficiency, and robustness in modern RL algorithms.

Focuses on cooperation and coordination challenges in multi-agent environments, highlighting real-world applications and inter-agent communication.

Provides a space for practitioners to share experiences and practical insights into implementing RL for complex real-world systems like supply chains, energy management, and power grids.

Bridges reinforcement learning and causal inference to enhance decision-making beyond standard (PO)MDPs, with a focus on generalization and credit assignment.

Advances beyond traditional reward-centric RL by exploring intrinsic motivation, skill discovery, and leveraging human-centric signals for building generalist agents.

Critically examines the conceptual foundations, axioms, and assumptions that frame reinforcement learning research problems.