NightFeats @ MMU-RAGent NeurIPS 2025: 面向文本到文本轨道的上下文优化多智能体RAG系统
NightFeats @ MMU-RAGent NeurIPS 2025: A Context-Optimized Multi-Agent RAG System for the Text-to-Text Track
AI总结：
提出一种结构化多智能体RAG系统NightFeats，通过检索、策展和组合三阶段分解知识合成，引入时序语义重排序、矛盾协调和引用保留架构，在MMU-RAGent竞赛中超越商业基线。
摘要：
我们提出NightFeats，一个结构化的多智能体检索增强生成（RAG）系统，提交至NeurIPS 2025的MMU-RAGent竞赛，并在文本到文本轨道中获得最佳动态评估奖。本文并非以基准最大化目标，而是提出一个原则性流水线，将知识合成为三个协调阶段：检索、策展和组合，每个阶段由显式的中间表示和交接契约控制。受智能体上下文工程（ACE）启发，该系统引入时序语义重排序、有界矛盾协调和保留引用的组合作为核心架构原语。竞赛结果表明，NightFeats在LLM-as-a-Judge和人类Likert评估中超越了包括Claude-SonnetV2和Nova-Pro在内的商业基线，证实了架构透明性和可验证证据基础比单纯优化自动相似度指标的系统更符合人类偏好。
英文摘要：
We present NightFeats, a structured multi-agent retrieval-augmented generation (RAG) system submitted to the MMU-RAGent competition at NeurIPS 2025, where it was awarded Best Dynamic Evaluation in the text-to-text track. Rather than targeting benchmark maximization, this work proposes a principled pipeline that decomposes knowledge synthesis into three coordinated phases: retrieval, curation, and composition, each governed by explicit intermediate representations and handoff contracts. Inspired by Agentic Context Engineering (ACE), the system introduces temporal-semantic reranking, bounded contradiction reconciliation, and citation-preserving composition as core architectural primitives. Competition results show that NightFeats surpasses proprietary baselines including Claude-SonnetV2 and Nova-Pro on LLM-as-a-Judge and Human Likert evaluations, confirming that architectural transparency and verifiable evidence grounding are better aligned with human preferences than systems optimizing narrowly for automatic similarity metrics.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
