Subtopic Deep Dive

Bayesian Optimal Experimental Design
Research Guide

What is Bayesian Optimal Experimental Design?

Bayesian Optimal Experimental Design selects experimental conditions to maximize expected utility under prior beliefs about model parameters using decision-theoretic criteria.

This approach integrates prior information into design criteria like Bayesian D-optimal or entropy reduction (Chaloner and Verdinelli, 1995, 1836 citations). It supports adaptive sequential designs for uncertain models. Over 10 key papers span reviews and applications from 1993 to 2020.

Curated Papers

Key Challenges

Why It Matters

Bayesian designs improve efficiency in high-cost experiments such as phase I cancer trials by optimizing dose escalation (Le Tourneau et al., 2009, 842 citations). In genomics, they control false discovery rates for gene expression analysis (Reiner-Benaim et al., 2003, 1763 citations). These methods enable personalized medicine and adaptive testing in agriculture (Smith et al., 2005, 594 citations), reducing sample sizes while enhancing inference precision.

Key Research Challenges

High-Dimensional Priors

Computing expected utilities over high-dimensional parameter spaces requires approximations like Laplace methods or MCMC. Chaloner and Verdinelli (1995) highlight analytical intractability for complex priors. This limits scalability in large-scale experiments.

Adaptive Sequential Design

Sequential designs update priors after each experiment, demanding real-time optimization. Greenhill et al. (2020, 450 citations) review Bayesian optimization for this but note computational expense. Balancing exploration and exploitation remains difficult.

Utility Specification

Defining utilities incorporating model uncertainty and decision goals is subjective. Schönbrodt and Wagenmakers (2017, 731 citations) discuss Bayes factor criteria for evidence planning. Misspecification leads to suboptimal designs.

Essential Papers

Bayesian Experimental Design: A Review

Kathryn Chaloner, Isabella Verdinelli · 1995 · Statistical Science · 1.8K citations

This paper reviews the literature on Bayesian experimental design. A unified view of this topic is presented, based on a decision-theoretic approach. This framework casts criteria from the Bayesian...

Identifying differentially expressed genes usingfalse discovery rate controlling procedures

Anat Reiner‐Benaim, Daniel Yekutieli, Yoav Benjamini · 2003 · Bioinformatics · 1.8K citations

Abstract Motivation: DNA microarrays have recently been used for the purpose of monitoring expression levels of thousands of genes simultaneously and identifying those genes that are differentially...

Optimum Experimental Designs

Robert E. Wheeler, Anthony C. Atkinson, A N Donev · 1993 · Technometrics · 1.5K citations

Part I. Fundamentals Introduction Some key ideas Experimental strategies The choice of a model Models and least squares Criteria for a good experiment Standard designs The analysis of experiments P...

Multiple Hypothesis Testing in Microarray Experiments

Sandrine Dudoit, Juliet Popper Shaffer, Jennifer C. Boldrick · 2003 · Statistical Science · 1.1K citations

DNA microarrays are part of a new and promising class of biotechnologies that allow the monitoring of expression levels in cells for thousands of genes simultaneously. An important and common quest...

Dose Escalation Methods in Phase I Cancer Clinical Trials

Christophe Le Tourneau, J. Jack Lee, Lillian L. Siu · 2009 · JNCI Journal of the National Cancer Institute · 842 citations

Phase I clinical trials are an essential step in the development of anticancer drugs. The main goal of these studies is to establish the recommended dose and/or schedule of new drugs or drug combin...

Bayes factor design analysis: Planning for compelling evidence

Felix D. Schönbrodt, Eric‐Jan Wagenmakers · 2017 · Psychonomic Bulletin & Review · 731 citations

Discussion of "Analysis of variance--why it is more important than ever" by A. Gelman

Gelman, Andrew · 2005 · arXiv (Cornell University) · 642 citations

Discussion of ``Analysis of variance--why it is more important than ever'' by A. Gelman [math.ST/0504499]

Reading Guide

Foundational Papers

Start with Chaloner and Verdinelli (1995) for decision-theoretic unification (1836 citations), then Atkinson et al. (1993) for optimum design theory (1465 citations).

Recent Advances

Study Greenhill et al. (2020) for Bayesian optimization review (450 citations) and Schönbrodt and Wagenmakers (2017) for Bayes factor planning (731 citations).

Core Methods

Expected utility maximization, posterior covariance minimization, mutual information criteria, Gaussian process approximations.

How PapersFlow Helps You Research Bayesian Optimal Experimental Design

Discover & Search

Research Agent uses searchPapers and exaSearch to find Chaloner and Verdinelli (1995) plus 50+ related works on Bayesian criteria; citationGraph reveals connections to Greenhill et al. (2020); findSimilarPapers expands to adaptive designs from Le Tourneau et al. (2009).

Analyze & Verify

Analysis Agent applies readPaperContent to extract utility formulas from Chaloner and Verdinelli (1995), verifies derivations via verifyResponse (CoVe), and runs PythonAnalysis with NumPy for simulating D-optimal designs; GRADE scores evidence strength on prior integration claims.

Synthesize & Write

Synthesis Agent detects gaps in sequential design coverage across papers, flags contradictions in utility definitions; Writing Agent uses latexEditText, latexSyncCitations for Chaloner (1995), and latexCompile to produce design criterion tables; exportMermaid visualizes decision-theoretic frameworks.

Use Cases

"Simulate Bayesian D-optimal design for dose escalation trial with normal prior."

Research Agent → searchPapers → Analysis Agent → runPythonAnalysis (NumPy/MCMC simulation) → matplotlib plot of expected utility vs. doses.

"Write LaTeX review comparing Bayesian vs. frequentist designs citing Chaloner 1995."

Synthesis Agent → gap detection → Writing Agent → latexEditText + latexSyncCitations + latexCompile → PDF with utility equations and bibliography.

"Find GitHub code for Bayesian optimization in experimental design from recent papers."

Research Agent → paperExtractUrls (Greenhill 2020) → Code Discovery → paperFindGithubRepo → githubRepoInspect → verified implementation notebooks.

Automated Workflows

Deep Research workflow scans 50+ papers from Chaloner (1995) to Greenhill (2020), producing structured reports on utility criteria evolution. DeepScan applies 7-step analysis with CoVe checkpoints to verify adaptive design claims in Le Tourneau et al. (2009). Theorizer generates novel utility functions from literature patterns in Schönbrodt and Wagenmakers (2017).

Try Doxa for Bayesian Optimal Experimental Design Research

Frequently Asked Questions

What defines Bayesian Optimal Experimental Design?

It maximizes expected posterior utility using priors, contrasting frequentist criteria like D-optimality (Chaloner and Verdinelli, 1995).

What are common methods?

Decision-theoretic utilities, entropy minimization, and Bayesian optimization via Gaussian processes (Greenhill et al., 2020).

What are key papers?

Chaloner and Verdinelli (1995, 1836 citations) provides the foundational review; Greenhill et al. (2020, 450 citations) covers recent adaptive methods.