One Set to Rule Them All: How to Obtain General Chemical Conditions via Bayesian Optimization over Curried Functions

2025-02-26

AI Summary

This work addresses the challenge of poor cross-task transferability of universal parameter sets in chemical reaction condition optimization. We propose a novel Bayesian optimization paradigm designed for generalization. Our core innovation models multi-task optimization as the joint optimization of curried functions, thereby decoupling task-specific and parameter spaces, and introduces a myopic strategy to coordinate exploration across the experimental space. Integrating Bayesian optimization, currying-based modeling, and multi-task joint search, our method is systematically evaluated on real chemical reaction datasets. Results demonstrate strong transferability of the learned universal conditions: high-performing parameter sets are identified with only a few cross-reaction experiments—matching the performance of sophisticated transfer learning methods while substantially reducing redundant tuning costs. This work establishes an interpretable, scalable framework for efficient, generalizable optimization of reaction conditions.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Optimization
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

General parameters are highly desirable in the natural sciences - e.g., chemical reaction conditions that enable high yields across a range of related transformations. This has a significant practical impact since those general parameters can be transferred to related tasks without the need for laborious and time-intensive re-optimization. While Bayesian optimization (BO) is widely applied to find optimal parameter sets for specific tasks, it has remained underused in experiment planning towards such general optima. In this work, we consider the real-world problem of condition optimization for chemical reactions to study how performing generality-oriented BO can accelerate the identification of general optima, and whether these optima also translate to unseen examples. This is achieved through a careful formulation of the problem as an optimization over curried functions, as well as systematic evaluations of generality-oriented strategies for optimization tasks on real-world experimental data. We find that for generality-oriented optimization, simple myopic optimization strategies that decouple parameter and task selection perform comparably to more complex ones, and that effective optimization is merely determined by an effective exploration of both parameter and task space.

Problem

• Optimize chemical reaction conditions
• Bayesian Optimization for general parameters
• Accelerate identification of general optima

Innovation

• Bayesian optimization over curried functions
• Generality-oriented optimization strategies
• Decoupling parameter and task selection