A Semi-supervised Molecular Learning Framework for Activity Cliff Estimation

TL;DR

This paper introduces SemiMol, a semi-supervised learning framework that improves molecular activity cliff prediction by leveraging unannotated data and a novel curriculum learning approach, outperforming existing methods especially in low-data scenarios.

Contribution

The paper presents SemiMol, a semi-supervised learning method with an instructor model and adaptive curriculum, specifically designed to handle activity cliffs in molecular property prediction.

Findings

SemiMol significantly improves prediction accuracy on activity cliff datasets.

It outperforms state-of-the-art pretraining and SSL baselines.

The approach is effective across 30 diverse datasets.

Abstract

Machine learning (ML) enables accurate and fast molecular property predictions, which are of interest in drug discovery and material design. Their success is based on the principle of similarity at its heart, assuming that similar molecules exhibit close properties. However, activity cliffs challenge this principle, and their presence leads to a sharp decline in the performance of existing ML algorithms, particularly graph-based methods. To overcome this obstacle under a low-data scenario, we propose a novel semi-supervised learning (SSL) method dubbed SemiMol, which employs predictions on numerous unannotated data as pseudo-signals for subsequent training. Specifically, we introduce an additional instructor model to evaluate the accuracy and trustworthiness of proxy labels because existing pseudo-labeling approaches require probabilistic outputs to reveal the model's confidence and fail to be applied in regression tasks. Moreover, we design a self-adaptive curriculum learning algorithm to progressively move the target model toward hard samples at a controllable pace. Extensive experiments on 30 activity cliff datasets demonstrate that SemiMol significantly enhances graph-based ML architectures and outpasses state-of-the-art pretraining and SSL baselines.