Self-driving laboratory: Science
Principles
The science it rests on, with why each principle matters to it
Relies on- Probability and statistics Many systems choose the next experiment by Bayesian optimisation [6], a method that uses probability to weigh which untried option is most likely to pay off.
as of 2026-11-11
The science explained
The science behind it explained for a curious reader
Imagine looking for the best spot on a huge dark hillside with only a torch. You cannot light every square metre, so you look at a few points, guess where the ground rises and walk that way, then look again. A self-driving laboratory does the same with recipes. The hillside is the space of possible experiments, more than 98 million of them in the Liverpool robot's search [4]. Each experiment is a flash of the torch. After each result the software updates its picture and picks the next point most likely to be higher [4]. Because every step uses what was learned before, the robot finds a good answer after trying only a tiny share of the options.
as of 2026-11-11
Sources
- Experiment too big? Hire a mobile robot scientist University of Liverpool, Cooper Group news, 2020
- Review of low-cost self-driving laboratories in chemistry and materials science: the 'frugal twin' concept S. Lo, S. G. Baird, J. Schrier and others (including B. Maruyama and A. Aspuru-Guzik), Royal Society of Chemistry, 2024