Knowledge-enhanced Relation Graph and Task Sampling for Few-shot Molecular Property Prediction

TL;DR

The paper tackles FSMPP by explicitly modeling the many-to-many relationships between molecules and properties with a knowledge-enhanced molecule–property multi-relation graph (MPMRG) that fuses substructure similarities (scaffolds, functional groups) with property information. It couples this graph with two task-sampling modules: a meta-training task sampler guided by task correlations and a policy-gradient auxiliary task sampler that selects highly related auxiliary properties, enabling efficient meta-knowledge learning. Empirical results on five FSMPP datasets show state-of-the-art ROC-AUC performance, including Tox21 reaching 87.62% in 10-shot and PCBA achieving substantial gains, validating the effectiveness of the graph-based relational modeling and relevance-aware task sampling. The approach offers a practical route to leverage fine-grained structural cues and task relationships to improve performance when labeled data are scarce, with implications for accelerated molecular discovery and property optimization.

Abstract

Recently, few-shot molecular property prediction (FSMPP) has garnered increasing attention. Despite impressive breakthroughs achieved by existing methods, they often overlook the inherent many-to-many relationships between molecules and properties, which limits their performance. For instance, similar substructures of molecules can inspire the exploration of new compounds. Additionally, the relationships between properties can be quantified, with high-related properties providing more information in exploring the target property than those low-related. To this end, this paper proposes a novel meta-learning FSMPP framework (KRGTS), which comprises the Knowledge-enhanced Relation Graph module and the Task Sampling module. The knowledge-enhanced relation graph module constructs the molecule-property multi-relation graph (MPMRG) to capture the many-to-many relationships between molecules and properties. The task sampling module includes a meta-training task sampler and an auxiliary task sampler, responsible for scheduling the meta-training process and sampling high-related auxiliary tasks, respectively, thereby achieving efficient meta-knowledge learning and reducing noise introduction. Empirically, extensive experiments on five datasets demonstrate the superiority of KRGTS over a variety of state-of-the-art methods. The code is available in https://github.com/Vencent-Won/KRGTS-public.

Figures (3)

• Figure 1: The pipeline of KRGTS in 2-way 1-shot setting.
• Figure 2: The comparison of molecular-property relation graph and knowledge-enhanced molecule-property graph.
• Figure 3: Experiments on the auxiliary task sampler.