Optimiser Repairs for Living-Cell Self-Driving Laboratories: Performance and Redundancy under Experimental Constraints

Optimiser repairs via exclusions and confirmation requests reproduce the baseline closed-loop trajectory exactly when they only target platform-forbidden conditions under permanent condition retirement.

• Why it matters: Living-cell self-driving labs must handle delayed readouts, shifting responses, and limited sample availability, but it is unclear when “repairs” help versus when constraints make them redundant.
• What they did: In simulations, the study proves exact baseline matching for evaluated pending-exclusion, previous-batch, and confirmation rules, then runs 33,900 policy simulations using 200–300 paired campaign seeds per setting.
• The result: Across 27 landscape, assay-size, and kernel settings, previous-batch exclusion shifts final normalised performance by -0.004 to +0.283 (Holm-adjusted significance in 13 settings), and all three interventions reproduce every saved baseline event on 200 of 200 paired seeds.

AI summary of the abstract · may contain errors · not medical advice · read the paper.