Quantifying Generalization in Reinforcement Learning

N/A

Quantifying Generalization in Reinforcement Learning Karl Cobbe 1 Oleg Klimov 1 Chris Hesse 1 Taehoon Kim 1 John Schulman 1 Abstract (Nichol et al., 2018), we seek to better quantify an agent’s In this paper, we investigate the problem of over- ability to generalize. fitting in deep reinforcement learning. Among To begin,…

saved by

• Sarah Pan

related reading

• Just Ask for Generalization | Eric Jangevjang.com
• [1709.06560] Deep Reinforcement Learning that Mattersarxiv.org
• Understanding deep learning requires rethinking generalizationarxiv.org
• 2203.00543arxiv.org
• [2602.05999] On the Role of Computation in Reinforcement Learningarxiv.org
• Q-learning is not yet scalableseohong.me
• [2605.01172] A Theory of Generalization in Deep Learningarxiv.org
• [2510.13786] The Art of Scaling Reinforcement Learning Compute for LLMsarxiv.org
• [2606.12016] Generalization Hacking: Models Can Game Reinforcement Learning by Preventing Behavioral Generalizationarxiv.org
• Teacher Algorithms for Deep RL Agents that Generalize in Procedurally Generated Environments – Developmental Systems, a Blog of the Flowers Labdevelopmentalsystems.org
• Reinforcement Learning in Newcomblike Problemsproceedings.neurips.cc
• [2201.02177] Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasetsarxiv.org