Reproduce “Asymptotic Consistency of Loss-Calibrated Variational Bayes” 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
Prateek Jaiswal, Harsha Honnappa, Vinayak A. Rao
The authors introduce loss-calibrated variational Bayes (LCVB), which lets a decision’s loss shape the approximate posterior used to choose an action. Unlike naive variational Bayes (NVB), it couples posterior approximation with decision selection. In a newsvendor simulation, median distance to the true best inventory decision falls as observations grow for both methods. Under stated conditions, both methods have asymptotically consistent decisions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.