用可验证谈判机制让推荐系统可靠满足政策约束。
PCN-Rec: Agentic Proof-Carrying Negotiation for Reliable Governance-Constrained Recommendation
- 分两代理协商：用户偏好与政策约束由不同智能体处理。
- 98.55%的合规通过率，推荐质量仅下降0.021 NDCG@10。
- 支持审计追踪，适合需合规的推荐场景。
基于大模型的推荐系统虽能生成高质量排序列表，却难以稳定满足治理约束（如长尾曝光最低要求或多样性）。本文提出PCN-Rec，一种证明携带式协商管道，将自然语言推理与确定性执行分离。基础推荐器（MF/CF）生成大小为W的候选窗口，由用户倡导者（优化相关性）与策略代理（强制约束）进行协商。中介大模型合成前N项推荐列表，并附带结构化证书（JSON）声明约束满足情况。确定性验证器从列表重新计算所有约束，仅接受验证通过的证书；若失败，则启动确定性贪婪修复生成合规列表并重新验证，形成可审计轨迹。在MovieLens-100K数据集上，对551名可行用户（W=80），PCN-Rec实现98.55%的通过率，相比无验证/修复的一次性单模型基线，仅造成0.021的NDCG@10绝对下降（0.403 vs. 0.424），差异具有统计显著性（p < 0.05）。
原文摘要 · Abstract (English)
Modern LLM-based recommenders can generate compelling ranked lists, but they struggle to reliably satisfy governance constraints such as minimum long-tail exposure or diversity requirements. We present PCN-Rec, a proof-carrying negotiation pipeline that separates natural-language reasoning from deterministic enforcement. A base recommender (MF/CF) produces a candidate window of size W, which is negotiated by two agents: a User Advocate optimizing relevance and a Policy Agent enforcing constraints. A mediator LLM synthesizes a top-N slate together with a structured certificate (JSON) describing the claimed constraint satisfaction. A deterministic verifier recomputes all constraints from the slate and accepts only verifier-checked certificates; if verification fails, a deterministic constrained-greedy repair produces a compliant slate for re-verification, yielding an auditable trace. On MovieLens-100K with governance constraints, PCN-Rec achieves a 98.55% pass rate on feasible users (n = 551, W = 80) versus a one-shot single-LLM baseline without verification/repair, while preserving utility with only a 0.021 absolute drop in NDCG@10 (0.403 vs. 0.424); differences are statistically significant (p < 0.05).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。