Reproduce “Active Inference Tree Search in Large POMDPs” 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
Domenico Maisto, Francesco Gregoretti, Karl Friston, Giovanni Pezzulo
Active Inference Tree Search (AcT) replaces exhaustive policy evaluation with selective tree search for planning under uncertainty. It guides trees using expected free energy, which scores preferred outcomes and information that reduces uncertainty, alongside a visit-count exploration prior. On a deceptive tree, AcT reaches deeper states faster than UCT and FE; this proof-of-principle result does not establish superiority in general.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.