Reproduce “Adaptive patch foraging in deep reinforcement learning agents” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Nathan J. Wispinski, Andrew Butcher, Kory W. Mathewson, Craig S. Chapman, Matthew M. Botvinick, Patrick M. Pilarski

The authors train recurrent deep reinforcement-learning agents to forage between refreshing patches in a 3D task, learning when to stay and travel. Agents stayed longer as patch distance increased. Their leaving times averaged 15.8 steps above the undiscounted marginal value theorem prediction, but differed by 0.9 steps from a discount-adjusted estimate, not significantly, nor does this establish optimality.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.