M34.6 CONNECT THE MECHANISM
Use models to propose experiments and test hypotheses
An AI system predicted 2.2 million new crystals. How many are new, real, useful materials? Work out hit rates and base rates, and follow a prediction all the way to the lab.
LESSON OVERVIEW14 min lesson
Lesson overview
An AI system predicted 2.2 million new crystals. How many are new, real, useful materials? Work out hit rates and base rates, and follow a prediction all the way to the lab.
What you’ll explore
- AI-assisted discovery combines hypothesis generation, experiment selection, measurement, and revision; novelty, causal evidence, reproducibility, and domain validation determine what has been discovered.
GO TO THE SOURCE
Original explanations, connected to the research.
Scaling deep learning for materials discovery (GNoME; Merchant et al., Nature 2023)An autonomous laboratory for the accelerated synthesis of inorganic materials (A-Lab; Szymanski et al., Nature 2023)Artificial Intelligence Driving Materials Discovery? Perspective on the Article: Scaling Deep Learning for Materials Discovery (Cheetham & Seshadri, Chemistry of Materials 2024)Challenges in High-Throughput Inorganic Materials Prediction and Autonomous Synthesis (Leeman et al., PRX Energy 2024)Mathematical discoveries from program search with large language models (FunSearch; Romera-Paredes et al., Nature 2023)A Tutorial on Bayesian OptimizationAI FeynmanSuggest a correction
A precise note can make an explanation better.
Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.