773: Deep Reinforcement Learning for Maximizing Profits, with Prof. Barrett Thomas
Super Data Science: ML & AI Podcast with Jon Krohn
00:00
Exploring Decision-Making Challenges with Neural Nets and Linear Approximation
This chapter discusses leveraging the Concord Solver to address combinatorial problems in MDPs with vast decision spaces, covering approaches like neural nets, Q-learning, linear approximation, and decision-making to handle routing challenges. It emphasizes the obstacles in training neural nets due to complex decision spaces and proposes utilizing linear approximation as a quicker solution, albeit with constraints.
Transcript
Play full episode
