Collaborative Contextual Bayesian Optimization

2026-04-20

AI Summary

This work addresses the challenge of jointly exploring context and design spaces in heterogeneous multi-client contextual Bayesian optimization by proposing a Collaborative Contextual Bayesian Optimization (CCBO) framework. CCBO introduces collaborative learning into contextual Bayesian optimization for the first time, supporting both online collaboration and offline initialization based on historical beliefs, with an optional privacy-preserving communication mechanism. By integrating Gaussian processes, multi-task learning, and Bayesian optimization, CCBO enables effective knowledge sharing and transfer across clients while maintaining high efficiency even under heterogeneity. Theoretical analysis establishes a sublinear regret bound for the proposed method, and empirical evaluations—including simulations and real-world hot-strip mill experiments—demonstrate that CCBO significantly outperforms existing approaches, exhibiting robust performance even in highly heterogeneous settings.

Technology Category

• Search and Optimization: Mixed Discrete/Continuous Search
• Reasoning under Uncertainty: Stochastic Optimization
• Machine Learning: Online Learning & Bandits

Application Category

• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Discovering optimal designs through sequential data collection is essential in many real-world applications. While Bayesian Optimization (BO) has achieved remarkable success in this setting, growing attention has recently turned to context-specific optimal design, formalized as Contextual Bayesian Optimization (CBO). Unlike BO, CBO is inherently more challenging as it must approximate an entire mapping from the context space to its corresponding optimal design, requiring simultaneous exploration across contexts and exploitation within each. In many modern applications, such tasks arise across multiple potentially heterogeneous but related clients, where collaboration can significantly improve learning efficiency. We propose CCBO, Collaborative Contextual Bayesian Optimization, a unified framework enabling multiple clients to jointly perform CBO with controllable contexts, supporting both online collaboration and offline initialization from peers' historical beliefs, with an optional privacy-preserving communication mechanism. We establish sublinear regret guarantees and demonstrate, through extensive simulations and a real-world hot rolling application, that CCBO achieves substantial improvements over existing approaches even under client heterogeneity. The code to reproduce the results can be found at https://github.com/cchihyu/Collaborative-Contextual-Bayesian-Optimization

Problem

• Contextual Bayesian Optimization
• Collaborative Optimization
• Optimal Design
• Client Heterogeneity
• Sequential Decision Making

Innovation

• Collaborative Bayesian Optimization
• Contextual Bayesian Optimization
• Federated Learning
• Sublinear Regret
• Privacy-Preserving Optimization