Autonomous materials discovery shows varied paths but same results
Divergent strategies and convergent outcomes in autonomous materials discovery
Artificial Intelligence
Summary
Finding new materials often involves computers searching many possibilities. This study had several computer agents run independent searches for porous materials that store methane. Even though their search strategies were very different, they all found similar top materials. However, some errors appeared repeatedly due to shared data issues. This shows that multiple runs can confirm solid findings but also reveal common mistakes.
What this means in practice
- •For materials design teams: Use repeated autonomous searches to confirm stable material candidates for gas storage despite different search methods.
- •For data quality engineers: Implement independent audits in screening workflows to detect common-mode errors caused by incomplete or biased input data.
Authors
Jihan Kim
Abstract
Scientific agents are mostly evaluated on whether they complete tasks or recover known results; we instead study variation across repeated open-ended campaigns. Sixteen separately initialized sessions of one model-harness configuration received a frozen database of 12,499 metal-organic frameworks, a methane-storage objective, a pinned protocol and a one-week budget. Strategies diverged into four approaches spanning 100--5,000 screened structures, and eight built 2,253 hypothetical structures. Yet the agents recovered the same materials frontier near 200 cm^3/cm^3, and an independent calculation of the database's porous region found its nine best structures all among their reports. Enforced checks on half the agents raised fresh-run reproduction from one of eight to eight of eight but could not detectably improve conclusion validity, because fifteen of sixteen agents selected the same audit-excluded entry, an incomplete structure whose missing anions created artificial pore volume. Replicated agents thus reveal both robust conclusions and common-mode errors from shared inputs.