David Abel on the Science of Agency @ RLDM 2025
TalkRL: The Reinforcement Learning Podcast
00:00
Reinforcement Learning: Environments vs. Agents
This chapter explores the critical role of modeling environments in reinforcement learning, emphasizing the need for a more robust understanding of agents comparable to established frameworks like Markov Decision Processes. The discussion highlights the limitations of reward maximization in goal specification and considers alternative approaches for capturing complex tasks. With insights into recent advancements, the chapter also examines research methodologies and the importance of collaboration and mentorship in fostering growth within the RL community.
Transcript
Play full episode
