Massively Multitask Networks for Drug Discovery

TL;DR

This paper demonstrates that massively multitask neural networks trained on extensive biological data significantly outperform single-task models in drug discovery, highlighting the importance of data sharing and algorithmic advances.

Contribution

It introduces a large-scale multitask neural architecture for drug discovery, showing improved predictive accuracy with more data and tasks, and analyzing transferability limitations.

Findings

Multitask networks outperform single-task models.

Adding more data and tasks improves performance.

Limited transferability to unseen tasks.

Abstract

Massively multitask neural architectures provide a learning framework for drug discovery that synthesizes information from many distinct biological sources. To train these architectures at scale, we gather large amounts of data from public sources to create a dataset of nearly 40 million measurements across more than 200 biological targets. We investigate several aspects of the multitask framework by performing a series of empirical studies and obtain some interesting results: (1) massively multitask networks obtain predictive accuracies significantly better than single-task methods, (2) the predictive power of multitask networks improves as additional tasks and data are added, (3) the total amount of data and the total number of tasks both contribute significantly to multitask improvement, and (4) multitask networks afford limited transferability to tasks not in the training set. Our results underscore the need for greater data sharing and further algorithmic innovation to accelerate the drug discovery process.