Newest papers

Conservative Meta-Agent Search: Architecture Adaptation under Non-Stationarity

Ali Alzahrani · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31436 · cited by 0 works (OpenAlex) · open-access PDF

Korean

자동화 된 토큰 업데이트는 더 낮은 대응 시스템 디자인은 일반적으로 즉각적인 스카일러 보상으로 정적 작업 스위트에서 평가됩니다. 배포 된 시스템은 지속적인 변화, 지연 된 의존적 피드백, 상관 관계 구성 요소 실패 및 비용이 많은 아키텍처 변화에 직면합니다. 우리는 정렬 된 메타 에이전트 아키텍처 검색을 정렬 된 임시 요소 대체로 평가합니다. 전환 비용과 어려운 위험 제약과 함께 프로모션은 보수적 인 메타 에이전트 검색 (CMAS)을 도입합니다. CMAS는 사전 지정된 마진 위에서 보수적 인 메타 에이전트 검색을 필요로합니다. CMAS는 파일화 된 아키텍처를 복구하고 인쇄 된 그래픽 편집을 제안

Machine translation (Meta M2M100); the English original follows.

Abstract

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 interventional component credit. Promotion requires an empirically calibrated, variance-penalized simultaneous score above a pre-specified margin, and is followed by a simulated canary stage and pre-specified rollback. Since backtests cannot certify meta-level skill, we introduce MetaInvest-Bench, a regime-switching simulator whose specified worker competence makes a surrogate-optimal architecture computable, with leakage-controlled replay, adversarial governance tracks, and five negative controls. Financial streams serve as a risk-sensitive testbed for meta-level generalization; we claim no alpha. Across 96 streams, CMAS cuts normalized mean surrogate dynamic-oracle regret from 0.116 to 0.104 and raises held-out utility from 0.44 to 0.48 over the strongest adaptive baseline; when extra workers add no value, it contracts to near-single-agent teams. Its deployment inference cost is ~6% lower, though its search-inclusive token cost is the highest of the methods compared. The observed per-update false-promotion rate is 3.4% against a nominal 5% target, and 4.8% in a sealed rerun. The calibration is empirical: it concerns an update-level replay statistic on an author-designed benchmark, and guarantees neither full-utility improvement, nor absence of simulated canary-stage harm, nor sequence-level error control.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Architecture, Testbed, Component (thermodynamics), Revenue, Regret

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange