Enhancing Activity Prediction Models in Drug Discovery with the Ability to Understand Human Language

TL;DR

This paper introduces CLAMP, a novel activity prediction model that leverages natural language understanding and a modular architecture to improve zero- and few-shot learning in drug discovery tasks.

Contribution

The paper proposes a new modular architecture with a contrastive pre-training objective enabling activity prediction models to adapt to new tasks through textual understanding without additional training.

Findings

CLAMP outperforms existing models on few-shot benchmarks.

CLAMP demonstrates strong zero-shot prediction capabilities.

The modular design and pre-training improve adaptability and accuracy.

Abstract

Activity and property prediction models are the central workhorses in drug discovery and materials sciences, but currently they have to be trained or fine-tuned for new tasks. Without training or fine-tuning, scientific language models could be used for such low-data tasks through their announced zero- and few-shot capabilities. However, their predictive quality at activity prediction is lacking. In this work, we envision a novel type of activity prediction model that is able to adapt to new prediction tasks at inference time, via understanding textual information describing the task. To this end, we propose a new architecture with separate modules for chemical and natural language inputs, and a contrastive pre-training objective on data from large biochemical databases. In extensive experiments, we show that our method CLAMP yields improved predictive performance on few-shot learning benchmarks and zero-shot problems in drug discovery. We attribute the advances of our method to the modularized architecture and to our pre-training objective.