GLA: Grounding Large Language Models in Molecular Hierarchy for Chemical Understanding

ACL ARR 2026 January Submission10040 Authors

Keywords: Multimodal applications, Molecular applications

Abstract: Conventional Euclidean geometries lead to structural distortion and entangle core pharmacophoric identities with peripheral groups. Existing molecule-language models, relying on linear or uniform encodings, often obscure the hierarchical organization of chemical semantics. To address this, we propose Geometric-Language Alignment (GLA), a framework integrating intrinsic molecular topology into large language models. GLA employs a mixed-curvature encoder that maps scaffolds to hyperbolic space to capture hierarchy, while encoding side-chains in Euclidean space. These representations are aligned with text via a dual-view contrastive objective and injected into a frozen language model. Experiments on cross-modal retrieval, captioning, and property prediction benchmarks show GLA consistently improves performance over baselines, suggesting that modeling geometric heterogeneity enhances the grounding between molecular structure and chemical language.

Paper Type: Long

Research Area: NLP Applications

Research Area Keywords: NLP Applications

Contribution Types: Publicly available software and/or pre-trained models

Languages Studied: English

Submission Number: 10040