Learning from learning machines: a new generation of AI technology to meet the needs of science

Abstract

We outline emerging opportunities and challenges to enhance the utility of AI for scientific discovery. The distinct goals of AI for industry versus the goals of AI for science create tension between identifying patterns in data versus discovering patterns in the world from data. If we address the fundamental challenges associated with "bridging the gap" between domain-driven scientific models and data-driven AI learning machines, then we expect that these AI models can transform hypothesis generation, scientific discovery, and the scientific process itself.

References in corpus (8)

• Distilling the Knowledge in a Neural Network
• Methods for Interpreting and Understanding Deep Neural Networks
• Think Locally, Act Locally: The Detection of Small, Medium-Sized, and Large Communities in Large Networks
• Local polynomial regression on unknown manifolds
• Characterizing possible failure modes in physics-informed neural networks
• The Scientific Method in the Science of Machine Learning
• Decoupling feature extraction from policy learning: assessing benefits of state representation learning in goal based robotics
• Interpreting Convolutional Neural Networks Through Compression