APSIPA Transactions on Signal and Information Processing > Vol 14 > Issue 2

Molecular Representation Learning via Hierarchical Graph Transformer

Zehua Wang, Wangxuan Institute of Computer Technology and Academy for Advanced Interdisciplinary Studies, Peking University, China, Yang Liu, Wangxuan Institute of Computer Technology, Peking University, China, Wei Hu, Wangxuan Institute of Computer Technology, Peking University, China, forhuwei@pku.edu.cn
 
Suggested Citation
Zehua Wang, Yang Liu and Wei Hu (2025), "Molecular Representation Learning via Hierarchical Graph Transformer", APSIPA Transactions on Signal and Information Processing: Vol. 14: No. 2, e107. http://dx.doi.org/10.1561/116.20240082

Publication Date: 23 Apr 2025
© 2025 Z. Wang, Y. Liu and W. Hu
 
Subjects
 

Share

Open Access

This is published under the terms of CC BY-NC.

Downloaded: 50 times

In this article:
Introduction 
Related Works 
Preliminary 
The Proposed Hierarchical Graph Transformer 
Experiments 
Conclusion 
References 

Abstract

Molecular Representation Learning (MRL) is widely applied in various downstream tasks, such as molecule generation, molecular property prediction and reaction prediction. Nevertheless, MRL faces several challenges posed by the vast chemical space and limited labeled-data availability. In this paper, we propose Hierarchical Graph Transformer (HieGT), integrating atom-level and motif-level representations to capture local-global characteristics of molecules over a hierarchical graph. Leveraging 2D topological and 3D geometric encoding, HieGT enhances intrinsic representation understanding of molecules. The proposed method achieves the state-of-the-art performance over the molecular property prediction dataset PCBA of Open Graph Benchmark (OGB), and competitive results on PCQM4Mv2 with better interpretability.

DOI:10.1561/116.20240082

Companion

APSIPA Transactions on Signal and Information Processing Special Issue - Three-dimensional Point Cloud Data Modeling, Processing, and Analysis
See the other articles that are part of this special issue.