arXiv cs.AI· Stuart Farmer·· 6 天前AI 评分22
EliteTable：面向可执行控制策略的自改进程序搜索
EliteTable: Self-Improving Program Search for Executable Control Policies
AI 导读
EliteTable 是一种由 LLM 引导的自改进程序搜索算法，用固定容量的高分方案表来指导新提案、已有方案编辑与多方案重组。在六个 Gymnasium 控制任务上各进行十代搜索、每代五十个候选槽位，六项任务的候选分数中位数均从第一代到末代上升，五项任务的选中优胜者搜索分数中位数上升；CartPole-v1 中十次搜索的优胜者均在第一代达到搜索分数上限。
正文 · 原文
Abstract:EliteTable is a self-improving program-search algorithm guided by large language models (LLMs) that uses a fixed-capacity table of high-scoring solutions to guide new proposals, edits of existing solutions, and remixes of several solutions. Performance was evaluated across six Gymnasium control tasks, each with ten generations and fifty candidate slots per generation. Each task was evaluated across ten independently initialized searches. Candidate-score medians rose between the first and final generations in all six tasks, and selected-winner search-score medians rose in five. In CartPole-v1, the selected winner reached the search-score ceiling in the first generation of all ten searches. All ten paired held-out endpoints improved in four of the six tasks. BipedalWalker-v3 and CarRacing-v3 improved without reaching their reporting targets. Six of ten Acrobot-v1 searches declined on held-out episodes, and the median paired change was negative. The results document improvement during program search and the transfer of selected-winner gains to new episodes in some tasks. Comparisons and ablations are needed to establish EliteTable's relative performance and the contributions of its components.
| Subjects: | Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE) |
| Cite as: | arXiv:2610.12536 [cs.AI] |
| (or arXiv:2610.12536v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.12536 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Stuart Farmer [view email]
[v1]
Thu, 8 Oct 2026 13:24:21 UTC (762 KB)
来源：arXiv cs.AI · arxiv.org