pith. sign in

arxiv: 2305.18435 · v2 · pith:V4EMCABFnew · submitted 2023-05-29 · 💻 cs.LG · stat.ME

Statistically Efficient Bayesian Sequential Experiment Design via Reinforcement Learning with Cross-Entropy Estimators

classification 💻 cs.LG stat.ME
keywords designdistributionlearningamortisedcross-entropyestimatorsmodelpolicies
0
0 comments X
read the original abstract

Reinforcement learning can learn amortised design policies for designing sequences of experiments. However, current amortised methods rely on estimators of expected information gain (EIG) that require an exponential number of samples on the magnitude of the EIG to achieve an unbiased estimation. We propose the use of an alternative estimator based on the cross-entropy of the joint model distribution and a flexible proposal distribution. This proposal distribution approximates the true posterior of the model parameters given the experimental history and the design policy. Our method overcomes the exponential-sample complexity of previous approaches and provide more accurate estimates of high EIG values. More importantly, it allows learning of superior design policies, and is compatible with continuous and discrete design spaces, non-differentiable likelihoods and even implicit probabilistic models.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Efficient Adaptive Data Acquisition via Pretrained Belief Representations

    cs.LG 2026-06 unverdicted novelty 6.0

    POLAR uses pretrained predictive foundation models as fixed belief-state encoders and trains only a lightweight policy head on top for amortised Bayesian experimental design, optimisation, and active learning.

  2. Constrained Bayesian Experimental Design via Online Planning

    stat.ML 2026-05 unverdicted novelty 6.0

    Combines offline amortized pre-training with online scenario-tree planning to optimize constrained Bayesian experimental designs, producing more informative sequences than prior methods.