Tree Bandits for Generative Bayes

arXiv:2404.10436v1 Announce Type: new
Abstract: In generative models with obscured likelihood, Approximate Bayesian Computation (ABC) is often the tool of last resort for inference. However, ABC demands many prior parameter trials to keep only a small fraction that passes an acceptance test. To accelerate ABC rejection sampling, this paper develops a self-aware framework that learns from past trials and errors. We apply recursive partitioning classifiers on the ABC lookup table to sequentially refine high-likelihood regions into boxes. Each box is …

abstract arxiv bayes bayesian computation cs.lg framework generative generative models however inference likelihood paper prior sampling small stat.co stat.me test tool tree type