RLC 2026
RL Beyond Rewards: Ingredients for Developing Generalist Agents
Yingchen Xu, Siddhant Agarwal, Pranaya Jajoo, Harshit Sikchi, Chuning Zhu, Abhishek Gupta, Amy Zhang, Caleb Chuck
About This Workshop
RL has traditionally focused on maximizing rewards, but intelligent agents often rely on reward-free interactions and diverse environmental signals. This workshop seeks to advance beyond traditional reward-centric RL by exploring intrinsic motivation, skill discovery, predictive and contrastive representation learning, and human-centric signals.
Building upon recent progress including foundational models employing scalable alternative signals, the workshop bridges theoretical insights and practical applications toward creating more versatile, adaptive decision-making agents.