Automating Science using Robot Scientists
summary
Robot scientists Adam and Eve execute full hypothesis cycles: Adam uses logic in Prolog and abductive inference over a directed labeled hypergraph of yeast metabolism, chooses experiments with a Bayesian cost model and active learning, controls a 5 m by 3 m by 3 m lab robot, and experimentally confirms novel gene–enzyme assignments—claimed as the first machine to autonomously discover scientific knowledge. Eve uses QSAR induction, active learning, and acoustic liquid handling to screen libraries and repurpose triclosan against Plasmodium vivax. Genesis targets 10,000 experiments/day in parallel for yeast systems biology.
AI-written summary · about
chapters (11)
- 0:00Intro
- 2:51Scientific Discovery
- 7:17The Concept of a Robot Scientist Computer system capable of originating its own experiments, physically executing them, interpreting the results, and then repeating the cycle
- 9:30Motivation: Technological Robot Scientists have the potential to increase the productivity of science. They can work cheaper, faster, more accurately, and longer than humans. They can also be easily multiplied. - Enabling the high-throughput festing of hypotheses.
- 11:38The Application Domain
- 16:08Genome Scale Model of Yeast Metabolism
- 21:39The Form of the Experiments
- 22:52Inferring Experiments Given a set of hypotheses we wish to infer an experiment that
- 26:26Formalising the Problem Use graphs and standard chemoinformatic methods to represent background knowledge - the use of relations is planned
- 30:41The Complexity of Biological System
- 31:38Genesis: 3rd Generation Robot Scientist
related talks
32mThe AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic...
17mBeyond Automation: AI and the Evolution of Robotic Intelligence
22mThe Nexus of Innovation: How Knowledge Graphs Fuel AI Scientific Breakthroughs
48mReprogramming Biology
19mReal-time Experiments with an AI Co-Scientist
49mPedro Domingos, Professor, University of Washington
19mAutonomous Agents for Scientific Tasks
59mMaking the most of artificial and human intelligence for data science