arXiv：cs.CL· Preeti Saraswat, Divya Neelagiri, Ajay Manoj·· 3 小时前
Gated Memory：面向对话式 AI 的准入控制记忆形成框架
Gated Memory: Admission-Controlled Memory Formation for Conversational AI
AI 导读
针对个性化对话 AI 记忆形成阶段缺乏原则性设计的问题，研究者提出 Gated Memory 框架，在对话与存储之间设置准入闸门与条件化富化两个决策检查点。该框架在抽取前依据完整话语上下文评估候选事实，并对准入内容做原子事实分解、来源标注与时间地点锚定。在 LoCoMo-10 基准上，其在检索与生成完全一致的前提下，LLM-judge 准确率相对强基线提升 2.6%。
正文
Abstract:Personalized conversational AI relies on long-term memory systems that extract facts from user utterances and store them in persistent vector stores. Despite progress in retrieval, deduplication, and lifecycle management, the formation stage, the moment a fact is first written to storage has received almost no principled attention. We identify this as the binding constraint on memory quality in production systems. Critical contextual signals, such as the distinction between a permanent user attribute and a transient situation, exist only in the original utterance and are irreversibly lost the moment extraction produces a subject-relation-object triple. No downstream process can recover them. We propose Gated Memory, a lightweight, modular formation framework that interposes two decision checkpoints between conversation and storage: an admission gate that evaluates every candidate fact against the full utterance context before extraction runs, and a conditional enrichment stage that grounds admitted facts through an entity scope taxonomy with privacy constraints. The gate evaluates only the current exchange while using prior turns as read-only reference context, and produces a structured formation record. Admitted content is decomposed into atomic facts, each categorized, tagged with provenance (directly stated versus inferred), scoped to its condition of applicability, and grounded in resolved time and place, subject to a constraint that no entity absent from the context may be asserted. On the LoCoMo-10 benchmark with atypical emotional density in utterance data, Gated Memory achieves an overall +2.6% relative improvement in LLM-judge accuracy over a strong baseline with identical retrieval and generation, establishing formation quality as a measurable constraint on memory performance.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Information Retrieval (cs.IR); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.11270 [cs.CL] |
| (or arXiv:2610.11270v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.11270 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Preeti Saraswat [view email]
[v1]
Thu, 8 Oct 2026 05:27:02 UTC (363 KB)
来源：arXiv：cs.CL · arxiv.org