Aligning AI-driven discovery with human intuition

TL;DR

This paper introduces a new principle for AI-driven scientific modeling that produces representations aligned with human intuition, facilitating better collaboration and understanding without prior physical knowledge.

Contribution

The paper proposes a novel general principle for distilling AI representations that are more interpretable and aligned with human scientific intuition, independent of prior physical knowledge.

Findings

AI-generated variables resemble those chosen by human scientists

Approach improves interpretability of AI models in scientific discovery

Method enhances human-AI collaboration in modeling processes

Abstract

As data-driven modeling of physical dynamical systems becomes more prevalent, a new challenge is emerging: making these models more compatible and aligned with existing human knowledge. AI-driven scientific modeling processes typically begin with identifying hidden state variables, then deriving governing equations, followed by predicting and analyzing future behaviors. The critical initial step of identification of an appropriate set of state variables remains challenging for two reasons. First, finding a compact set of meaningfully predictive variables is mathematically difficult and under-defined. A second reason is that variables found often lack physical significance, and are therefore difficult for human scientists to interpret. We propose a new general principle for distilling representations that are naturally more aligned with human intuition, without relying on prior physical knowledge. We demonstrate our approach on a number of experimental and simulated system where the variables generated by the AI closely resemble those chosen independently by human scientists. We suggest that this principle can help make human-AI collaboration more fruitful, as well as shed light on how humans make scientific modeling choices.