Learning Molecular Chirality via Chiral Determinant Kernels

2026-02-07

AI Summary

Existing molecular representation methods struggle to jointly model central and axial chirality, limiting the application of machine learning in stereochemical tasks. This work proposes ChiDeK, a novel framework that explicitly encodes both types of chirality within a unified architecture for the first time. It constructs an SE(3)-invariant chiral matrix using a chiral determinant kernel and integrates local chiral information into global 3D molecular graph representations via a cross-attention mechanism. The study also introduces the first benchmark dataset tailored for axial chirality, featuring electronic circular dichroism (ECD) spectra and optical rotation predictions. ChiDeK significantly outperforms existing methods across four tasks—R/S configuration classification, enantiomer ranking, ECD spectrum prediction, and optical rotation estimation—achieving an average accuracy improvement of over 7% on axial chirality tasks.

Technology Category

• Machine Learning: Representation Learning
• Knowledge Representation and Reasoning: Other Foundations of Knowledge Representation & Reasoning
• Computer Vision: Representation Learning for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Chirality is a fundamental molecular property that governs stereospecific behavior in chemistry and biology. Capturing chirality in machine learning models remains challenging due to the geometric complexity of stereochemical relationships and the limitations of traditional molecular representations that often lack explicit stereochemical encoding. Existing approaches to chiral molecular representation primarily focus on central chirality, relying on handcrafted stereochemical tags or limited 3D encodings, and thus fail to generalize to more complex forms such as axial chirality. In this work, we introduce ChiDeK (Chiral Determinant Kernels), a framework that systematically integrates stereogenic information into molecular representation learning. We propose the chiral determinant kernel to encode the SE(3)-invariant chirality matrix and employ cross-attention to integrate stereochemical information from local chiral centers into the global molecular representation. This design enables explicit modeling of chiral-related features within a unified architecture, capable of jointly encoding central and axial chirality. To support the evaluation of axial chirality, we construct a new benchmark for electronic circular dichroism (ECD) and optical rotation (OR) prediction. Across four tasks, including R/S configuration classification, enantiomer ranking, ECD spectrum prediction, and OR prediction, ChiDeK achieves substantial improvements over state-of-the-art baselines, most notably yielding over 7% higher accuracy on axially chiral tasks on average.

Problem

• molecular chirality
• axial chirality
• stereochemical representation
• machine learning
• chirality encoding

Innovation

• chiral determinant kernel
• axial chirality
• SE(3)-invariant representation
• cross-attention
• molecular representation learning