Reproduce “Decomposition Strategies for Constructive Preference Elicitation” 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
Paolo Dragone, Stefano Teso, Mohit Kumar, Andrea Passerini
PCL decomposes constructive preference elicitation into part-wise inference and feedback, holding other parts fixed and providing relevant context. It learns feature-based utility weights through perceptron-style updates while recombining parts into complete configurations. In simulated tasks, hotel regret fell by an order of magnitude in around 100 iterations. The method retains local, not global, optimality guarantees under conditionally alpha-informative feedback assumptions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.