Cap learns policies to optimize experiments from limited task data
CAP: Learning to Optimize from Sparse Observations of Related Tasks
Machine Learning
Summary
Running experiments to find the best solution can be costly and slow, especially when there are few tries allowed. The authors propose CAP, a method that learns from previous similar tasks to make better decisions on which experiments to run next. CAP uses past data to simulate practice tasks and trains itself to pick the best next step, saving resources when working on new problems. Their tests show CAP works well on both synthetic tasks and real engineering simulations, meaning it helps get better results faster with fewer costly tries.
What this means in practice
- •For industrial process engineers: Optimize complex processes by selecting experiments more effectively using prior related optimization data.
- •For machine learning system builders: Develop adaptive query strategies that improve optimization using limited data from related problem instances.
Authors
Lin Fu, Siwei Fu, Dong Ni
Abstract
In experiment-driven black-box optimization, each evaluation can consume materials, equipment time, and other resources, limiting the number of experiments available for a new task. Records from completed related optimization campaigns could help, but learning a sequential query policy remains challenging when task-specific experience consists of only a few sparse histories. We propose CAP (Candidate Acquisition Policy), which models shared structure and task-specific variation in fixed historical records to sample interactive practice functions. On these functions, CAP learns a feedback-conditioned query policy aimed at final optimization performance under a fixed evaluation budget. At deployment, its parameters remain fixed while each new observation informs the next query. Across five synthetic task families spanning 6 and 20 dimensions and two engineering simulation benchmarks, BOPTEST and SU2, CAP attains the lowest median final-budget value among the reported methods on four synthetic families and both engineering benchmarks. These results indicate that sparse records from completed related tasks can support adaptive query decisions on new tasks, making more effective use of limited, costly experimental evaluations.