2016/12/04 by Rudy Bunel, Alban Desmaison, Bunel, Rudy +8
Computer Science · Decision Sciences · #Advanced Bandit Algorithms Research #Evolutionary Algorithms and Applications #FOS: Computer and information sciences #Machine Learning (cs.LG) #Metaheuristic Optimization Algorithms Research #cs.LG
paper · pdf · doi:10.48550/arxiv.1612.01094
Workshop version for the NIPS NAMPI Workshop. Extended version at arXiv:1611.01787
arxiv created 2016/12/04 · openalex publication_date 2016/12/04 · arxiv updated 2016/12/06 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Superoptimization requires the estimation of the best program for a given computational task. In order to deal with large programs, superoptimization techniques perform a stochastic search. This involves proposing a modification of the current program, which is accepted or rejected based on the improvement achieved. The state of the art method uses uniform proposal distributions, which fails to exploit the problem structure to the fullest. To alleviate this deficiency, we learn a proposal distribution over possible modifications using Reinforcement Learning. We provide convincing results on the superoptimization of "Hacker's Delight" programs.