Benchmarking the Discovery Engine

TL;DR

This paper introduces the Discovery Engine, an automated system that combines machine learning and interpretability to enhance scientific discovery across various fields, outperforming previous methods in predictive accuracy and insight generation.

Contribution

The paper presents the Discovery Engine as a new benchmark tool that improves predictive performance and interpretability in scientific data analysis across multiple disciplines.

Findings

Matches or exceeds prior predictive performance

Provides deeper, more actionable insights

Demonstrates potential as a new standard for scientific modeling

Abstract

The Discovery Engine is a general purpose automated system for scientific discovery, which combines machine learning with state-of-the-art ML interpretability to enable rapid and robust scientific insight across diverse datasets. In this paper, we benchmark the Discovery Engine against five recent peer-reviewed scientific publications applying machine learning across medicine, materials science, social science, and environmental science. In each case, the Discovery Engine matches or exceeds prior predictive performance while also generating deeper, more actionable insights through rich interpretability artefacts. These results demonstrate its potential as a new standard for automated, interpretable scientific modelling that enables complex knowledge discovery from data.