Uncertainty-Aware Multi-Objective Reinforcement Learning-Guided Diffusion Models for 3D De Novo Molecular Design

2025-10-24

AI Summary

Generating high-quality, three-dimensional (3D) drug-like molecules under multiple, often conflicting, objectives—such as potency, ADMET properties, and structural stability—remains a major challenge in de novo drug discovery. Method: This paper proposes a novel generative framework integrating diffusion modeling with multi-objective reinforcement learning. It introduces an uncertainty-aware surrogate model to dynamically modulate the reward function, enabling adaptive trade-off optimization across molecular properties. Furthermore, closed-loop feedback is incorporated via molecular dynamics simulations and ADMET prediction to enhance pharmacological relevance and conformational stability. Contribution/Results: Evaluated on multiple benchmark datasets, our approach significantly outperforms state-of-the-art baselines. Generated molecules exhibit high structural diversity, superior ADMET profiles, and target-binding stability comparable to clinical EGFR inhibitors. The framework establishes a new paradigm for controllable, physics-informed 3D molecular generation.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search
• Computer Vision: Diffusion Models for Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Designing de novo 3D molecules with desirable properties remains a fundamental challenge in drug discovery and molecular engineering. While diffusion models have demonstrated remarkable capabilities in generating high-quality 3D molecular structures, they often struggle to effectively control complex multi-objective constraints critical for real-world applications. In this study, we propose an uncertainty-aware Reinforcement Learning (RL) framework to guide the optimization of 3D molecular diffusion models toward multiple property objectives while enhancing the overall quality of the generated molecules. Our method leverages surrogate models with predictive uncertainty estimation to dynamically shape reward functions, facilitating balance across multiple optimization objectives. We comprehensively evaluate our framework across three benchmark datasets and multiple diffusion model architectures, consistently outperforming baselines for molecular quality and property optimization. Additionally, Molecular Dynamics (MD) simulations and ADMET profiling of top generated candidates indicate promising drug-like behavior and binding stability, comparable to known Epidermal Growth Factor Receptor (EGFR) inhibitors. Our results demonstrate the strong potential of RL-guided generative diffusion models for advancing automated molecular design.

Problem

• Optimizing 3D molecular structures for multiple property objectives
• Controlling multi-objective constraints in molecular diffusion models
• Enhancing molecular quality and drug-like behavior in design

Innovation

• Reinforcement Learning guides 3D molecular diffusion models
• Uncertainty-aware surrogate models dynamically shape rewards
• Multi-objective optimization balances molecular properties and quality