Reproduce “Inverse Reinforcement Learning with Switching Rewards and History Dependency for Characterizing Animal Behaviors” 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
Jingyang Ke, Feiyang Wu, Jiyi Wang, Jeffrey Markowitz, Anqi Wu
SWIRL infers changing goals in animal-behavior sequences by making mode switches depend on current state and actions on recent locations. In a water-restricted maze, it identified water-, home-, and explore-related modes across 238 trajectories from 10 mice. The full model had the strongest held-out log-likelihood among tested variants and a MaxEnt baseline; inferred rewards are not measurements of motivation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.