On scientific understanding with artificial intelligence

Abstract

Imagine an oracle that correctly predicts the outcome of every particle physics experiment, the products of every chemical reaction, or the function of every protein. Such an oracle would revolutionize science and technology as we know them. However, as scientists, we would not be satisfied with the oracle itself. We want more. We want to comprehend how the oracle conceived these predictions. This feat, denoted as scientific understanding, has frequently been recognized as the essential aim of science. Now, the ever-growing power of computers and artificial intelligence poses one ultimate question: How can advanced artificial systems contribute to scientific understanding or achieve it autonomously? We are convinced that this is not a mere technical question but lies at the core of science. Therefore, here we set out to answer where we are and where we can go from here. We first seek advice from the philosophy of science to understand scientific understanding. Then we review the current state of the art, both from literature and by collecting dozens of anecdotes from scientists about how they acquired new conceptual understanding with the help of computers. Those combined insights help us to define three dimensions of android-assisted scientific understanding: The android as a I) computational microscope, II) resource of inspiration and the ultimate, not yet existent III) agent of understanding. For each dimension, we explain new avenues to push beyond the status quo and unleash the full power of artificial intelligence's contribution to the central aim of science. We hope our perspective inspires and focuses research towards androids that get new scientific understanding and ultimately bring us closer to true artificial scientists.

13 pages, 3 figures, comments welcome!

References in corpus (6)

• Methods for Interpreting and Understanding Deep Neural Networks
• 11 TeraFLOPs per second photonic convolutional accelerator for deep learning optical neural networks
• A Survey on the Explainability of Supervised Machine Learning
• Entanglement by Path Identity
• Machine Learning of Explicit Order Parameters: From the Ising Model to SU(2) Lattice Gauge Theory
• Sonification and Sound Design for Astronomy Research, Education and Public Engagement

Cited by in corpus (37)

• Towards self-driving laboratories: The central role of density functional theory in the AI age
• 14 Examples of How LLMs Can Transform Materials Science and Chemistry: A Reflection on a Large Language Model Hackathon
• Artificial Intelligence and Machine Learning for Quantum Technologies
• A critical examination of robustness and generalizability of machine learning prediction of materials properties
• On the redundancy in large material datasets: efficient and robust learning with less data
• Predicting the Future of AI with AI: High-quality link prediction in an exponentially growing knowledge network
• Probing out-of-distribution generalization in machine learning for materials
• Perspectives on adaptive dynamical systems
• Asset management, condition monitoring and Digital Twins: damage detection and virtual inspection on a reinforced concrete bridge
• Toward Accurate Interpretable Predictions of Materials Properties within Transformer Language Models
• Designing Materials Acceleration Platforms for Heterogeneous CO2 Photo(thermal)catalysis
• Mechanical Metamaterials Fabricated from Self-assembly: A Perspective
• Digital Discovery of 100 diverse Quantum Experiments with PyTheus
• Towards a Benchmark for Scientific Understanding in Humans and Machines
• APACE: AlphaFold2 and advanced computing as a service for accelerated discovery in biophysics
• Forecasting high-impact research topics via machine learning on evolving knowledge graphs
• XLuminA: An Auto-differentiating Discovery Framework for Super-Resolution Microscopy
• Addressing the Non-perturbative Regime of the Quantum Anharmonic Oscillator by Physics-Informed Neural Networks
• Large Physics Models: Towards a collaborative approach with Large Language Models and Foundation Models
• Agents for self-driving laboratories applied to quantum computing
• Quantum Similarity Testing with Convolutional Neural Networks
• AI for Scientific Discovery is a Social Problem
• Autonomous microARPES
• Variational Optimization for Quantum Problems using Deep Generative Networks
• Virtual Reality for Understanding Artificial-Intelligence-driven Scientific Discovery with an Application in Quantum Optics
• Machine-learning the spectral function of a hole in a quantum antiferromagnet
• Meta-Designing Quantum Experiments with Language Models
• Experimental Solutions to the High-Dimensional Mean King's Problem
• Machine learning a time-local fluctuation theorem for nonequilibrium steady states
• Synergy between deep neural networks and the variational Monte Carlo method for small ${}^{4}H{e}_N$ clusters
• Digital Discovery of a Scientific Concept at the Core of Experimental Quantum Optics
• Digital Discovery of interferometric Gravitational Wave Detectors
• Deep Quantum Graph Dreaming: Deciphering Neural Network Insights into Quantum Experiments
• Learning the dynamics of Markovian open quantum systems from experimental data
• AI ensemble for signal detection of higher order gravitational wave modes of quasi-circular, spinning, non-precessing binary black hole mergers
• Detecting the Largest Correlations using the Correlation Density Matrix: a Quantum Monte Carlo Approach
• Automated Discovery of Non-local Photonic Gates