Multi-Objective Bayesian Optimization with Independent Tanimoto Kernel Gaussian Processes for Diverse Pareto Front Exploration

2025-08-12

AI Summary

To address the challenges of inefficient modeling of high-dimensional sparse molecular fingerprints and insufficient diversity of the Pareto front in multi-objective molecular optimization, this paper proposes GP-MOBO—a novel multi-objective Bayesian optimization algorithm. GP-MOBO is the first to integrate a lightweight exact Gaussian process with independent Tanimoto kernels, operating directly on the full molecular fingerprint space without dimensionality reduction and accepting SMILES strings as input. Its core innovation lies in balancing computational efficiency with modeling fidelity, thereby significantly enhancing the breadth and diversity of Pareto front exploration. On the DockSTRING benchmark, GP-MOBO achieves rapid convergence within only 20 iterations, attaining a higher geometric mean objective value and generating a larger number of high-quality, synthetically viable candidate molecules. These results demonstrate both its practicality under resource constraints and its state-of-the-art performance.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Optimization
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

We present GP-MOBO, a novel multi-objective Bayesian Optimization algorithm that advances the state-of-the-art in molecular optimization. Our approach integrates a fast minimal package for Exact Gaussian Processes (GPs) capable of efficiently handling the full dimensionality of sparse molecular fingerprints without the need for extensive computational resources. GP-MOBO consistently outperforms traditional methods like GP-BO by fully leveraging fingerprint dimensionality, leading to the identification of higher-quality and valid SMILES. Moreover, our model achieves a broader exploration of the chemical search space, as demonstrated by its superior proximity to the Pareto front in all tested scenarios. Empirical results from the DockSTRING dataset reveal that GP-MOBO yields higher geometric mean values across 20 Bayesian optimization iterations, underscoring its effectiveness and efficiency in addressing complex multi-objective optimization challenges with minimal computational overhead.

Problem

• Optimizing multiple molecular objectives efficiently
• Exploring diverse Pareto fronts in chemical space
• Reducing computational overhead in Bayesian optimization

Innovation

• Independent Tanimoto Kernel Gaussian Processes
• Efficient handling of full fingerprint dimensionality
• Broader chemical space exploration with Pareto proximity