Reinforcement learning is supervised learning on optimized data – The Berkeley Artificial Intelligence Research Blog

The BAIR Blog

The two most common perspectives on Reinforcement learning (RL) are optimization and dynamic programming . Methods that compute the gradients of the non-differentiable expected reward objective, such as the REINFORCE trick are commonly grouped into the optimization perspective, whereas methods that employ TD-learning or Q-learning are dynamic programming methods. While these methods have shown considerable success in recent years, these methods are still quite challenging to apply to new problems. In contrast deep supervised learning has been extremely successful and we may hence ask: Can we u

related reading

• A (Long) Peek into Reinforcement Learning | Lil'Loglilianweng.github.io
• ML Mentorship: Some Q/A about RLevjang.com
• Reward is not the optimization target — LessWronglesswrong.com
• Should I Use Offline RL or Imitation Learning? – The Berkeley Artificial Intelligence Research Blogbair.berkeley.edu
• Q-learning is not yet scalableseohong.me
• Reward Is Not the Optimization Targetturntrout.com
• Deep Reinforcement Learning: Pong from Pixelskarpathy.github.io
• Contrastive Learning as Goal-Conditioned Reinforcement Learningarxiv.org
• SFT, RL, and On-Policy Distillation Through a Distributional Lens | whnrehiew.github.io
• [2602.11399] Can We Really Learn One Representation to Optimize All Rewards?arxiv.org
• A Gallery of Methods Beyond RL — Part I: Sampling Methodsshengyu-feng.github.io
• An Updated Introduction to Reinforcement Learning | Sri's Blogsrianumakonda.com