BiMol-Diff: A Unified Diffusion Framework for Molecular Generation and Captioning

2026-04-27

AI Summary

This work proposes a unified diffusion framework for text-guided molecular generation and molecular captioning, addressing key challenges including the misalignment between molecular structures and natural language, insufficient long-range dependency modeling in autoregressive approaches, and structural distortion of critical substructures caused by uniform noise scheduling in standard diffusion models. The core innovation lies in a position-aware noise scheduling mechanism that dynamically modulates denoising intensity based on token reconstruction difficulty, thereby preserving essential molecular substructures. By integrating molecular graph representations with language encodings, the method achieves a 15.4% relative improvement in Exact Match for molecule reconstruction on ChEBI-20 and M3-20M benchmarks, while also attaining state-of-the-art performance in caption generation as measured by BLEU and BERTScore metrics.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

Bridging molecular structures and natural language is essential for controllable design. Autoregressive models struggle with long-range dependencies, while standard diffusion processes apply uniform corruption across positions, which can distort structurally informative tokens. We present BiMol-Diff, a unified diffusion framework for the paired tasks of text-conditioned molecule generation and molecule captioning. Our key component is a token-aware noise schedule that assigns position-dependent corruption based on token recovery difficulty, preserving harder-to-recover substructures during the forward process. On ChEBI-20 and M3-20M, BiMol-Diff improves molecule reconstruction with a 15.4% relative gain in Exact Match and achieves strong captioning results, attaining best BLEU and BERTScore among compared baselines. These results indicate token-aware noising improves fidelity in molecular structure-language modelling.

Problem

• molecular generation
• molecule captioning
• diffusion models
• structure-language modeling
• token-aware noise

Innovation

• token-aware noise schedule
• unified diffusion framework
• molecular generation
• molecule captioning
• structure-language modeling