Automatic Annotation Augmentation Boosts Translation between Molecules and Natural Language

2025-02-10

AI Summary

High-quality annotated data scarcity hinders molecular–language cross-modal modeling in biomedicine. Method: We propose LA³, a large language model (LLM)-driven automatic annotation augmentation framework, introducing a novel LLM-prompted annotation rewriting paradigm that generates semantically consistent, syntactically diverse, and lexically rich molecular descriptions—compatible with multimodal molecular representations (e.g., SMILES, InChI). Leveraging LA³, we construct LaChEBI-20, the first large-scale augmented molecular semantic dataset, and train LaMolT5—a T5-based molecular language model supporting molecular generation, description, and cross-modal (text/image/graph) transfer tasks. Results: LaMolT5 achieves state-of-the-art performance across multiple molecular language benchmarks, outperforming prior methods by up to 301% relative improvement. Ablation studies confirm the generalizability of our augmentation strategy across heterogeneous modalities.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Computer Vision: Language and Vision

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Recent advancements in AI for biological research focus on integrating molecular data with natural language to accelerate drug discovery. However, the scarcity of high-quality annotations limits progress in this area. This paper introduces LA$^3$, a Language-based Automatic Annotation Augmentation framework that leverages large language models to augment existing datasets, thereby improving AI training. We demonstrate the effectiveness of LA$^3$ by creating an enhanced dataset, LaChEBI-20, where we systematically rewrite the annotations of molecules from an established dataset. These rewritten annotations preserve essential molecular information while providing more varied sentence structures and vocabulary. Using LaChEBI-20, we train LaMolT5 based on a benchmark architecture to learn the mapping between molecular representations and augmented annotations. Experimental results on text-based *de novo* molecule generation and molecule captioning demonstrate that LaMolT5 outperforms state-of-the-art models. Notably, incorporating LA$^3$ leads to improvements of up to 301% over the benchmark architecture. Furthermore, we validate the effectiveness of LA$^3$ notable applications in *image*, *text* and *graph* tasks, affirming its versatility and utility.

Problem

• Enhances molecular-natural language translation
• Augments datasets for AI training
• Improves drug discovery through AI

Innovation

• LA$^3$ augments molecular datasets
• LaMolT5 enhances molecule-language mapping
• LA$^3$ improves AI training benchmarks