Anubis: Bayesian optimization with unknown feasibility constraints for scientific experimentation

2025/01/01 by Riley J. Hickman, Gary Tom, Yunheng Zou +2 · 1 voice · 1 citation
Computer Science · Mathematics · #Advanced Multi-Objective Optimization Algorithms #Statistics Education and Methodologies

paper · pdf · doi:10.1039/d5dd00018a

openalex publication_date 2025/01/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/01

Abstract

Model-based optimization strategies, such as Bayesian optimization (BO), have been deployed across the natural sciences in design and discovery campaigns due to their sample efficiency and flexibility.

Citations

• Variational Inference: A Review for Statisticians
• A Deep Learning Approach to Antibiotic Discovery
• Bayesian reaction optimization as a tool for chemical synthesis
• Atlas: a brain for self-driving laboratories
• Bayesian Optimization with Unknown Constraints
• Optimization Under Unknown Constraints
• BoFire: Bayesian Optimization Framework Intended for Real Experiments

Cited by

• Constrained composite Bayesian optimization for rational synthesis of polymeric particles

Discussions

• "In scientific discovery, not all experiments work out. Syntheses can fail, materials can be unstable. How can BO handle this? Anubis looks at BO with "unknown infeasibility." Feasibility-aware BO le [bsky, 1 points, 1 comments]

Related

• Honegumi: An Interface for Accelerating the Adoption of Bayesian Optimization in the Experimental Sciences
• BayBE: a Bayesian Back End for experimental planning in the low-to-no-data regime
• HypBO: Accelerating Black-Box Scientific Experiments Using Experts' Hypotheses
• How Useful is Intermittent, Asynchronous Expert Feedback for Bayesian Optimization?
• Optimizing the Unknown: Black Box Bayesian Optimization with Energy-Based Model and Reinforcement Learning