Ali Atiah Alzahrani · 2026 · arXiv
DOI: 10.48550/arXiv.2610.31436 · open-access PDF
自动代理系统设计通常在静态任务套件上进行评估,即时可扩展的奖励。部署的系统面临持续的转变,延迟的依赖反馈,相关的组件失败和昂贵的架构变化。我们将元代理架构搜索正式化为连续的现有替换,具有切换成本和硬风险限制,并引入保守的元代理搜索(CMAS)。
Machine translation (Meta M2M100); the English original follows.
Automated agentic-system design is typically evaluated on static task suites with immediate scalar rewards. Deployed systems face persistent shift, delayed dependent feedback, correlated component failures, and costly architecture changes. We formalize meta-agent architecture search as sequential incumbent replacement with switching costs and hard risk constraints and introduce Conservative Meta-Agent Search (CMAS). CMAS retrieves archived architectures and proposes typed graph edits, then scores each against the incumbent by paired shadow replay under common random numbers, with interventiona
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.31436, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange