Reproduce “Model selection for behavioral learning data and applications to contextual bandits” 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
Julien Aubert, Louis Köhler, Luc Lehéricy, Giulia Mezzadri, Patricia Reynaud-Bouret
The authors develop two methods to infer learning strategies from dependent choice sequences. Chronological hold-out compares candidate models; penalized likelihood jointly estimates parameters and selects among partition-based contextual-bandit models. In simulations, both have errors, though penalized likelihood selects OnePerItem correctly in more than 60% of cases. In real data, it assigns about 70% of 176 participants to that model, not proving participants’ strategy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.