paper-with-me

홈 › Papers

TRACE-Memory: Public-Conditioned Retrieval and Utility-Aware Evidence Admission for Personalized Generation

2026-08-09 · Jing Wang, Zhu Wang, Yifan Guo, Yulong Yang, Yunji Liang arxiv

Personalized generation systems retrieve user history by request--memory relevance and inject it into the model context. Yet relevant history may concern the wrong preference aspect, duplicate public information, or provide insufficient support. We argue that personal memory should be used only when it adds utility beyond a public-only response. We propose TRACE-Memory, a two-stage framework for selective personalization. Stage 1 queries for user-specific information missing from the request and public context, then retrieves a coverage-oriented candidate pool. Stage 2 admits a compact subset of source-traceable evidence units, or the empty set, according to response-level incremental utility. We progressively train the query-generation and evidence-admission policies through structured SFT initialization, reduced-space stage-wise GRPO warm-up, and nested multi-sample Joint GRPO. Across 4,500 Controlled and Natural tasks from Goodreads, Amazon Reviews, and Reddit, TRACE-Memory consistently outperforms random and lexical memory use, improves over semantic retrieval, remains competitive with frontier-LLM memory pipelines as local generator capacity increases, and conditions evidence admission on public-context sufficiency, supporting selective rather than default personalization.

📄 PDF Abstract BibTeX arXiv:2608.08446

Code (0)

등록된 구현이 없습니다.

Tasks

Semantic Retrieval

Similar Papers 제목 키워드 기반

Beyond Retrieval: Query-Conditioned Reuse of Long-Horizon Agent Trajectories

2026-08-13 · Yifei Li, Heng Wang, Lingling Zhang, Muye Huang 외 arxiv

Retrieval can identify a past trajectory that may matter, yet it does not specify how an acting agent should use that trajectory after users, entities, constraints, or environment state have changed. We identify this pos…

Narrative World Model: Narratology-Grounded Writer Memory for Long-Form Fiction

2026-07-06 · Mohammad Saifullah, Thomas Kornmaier, Taaha Kazi, Vasu Sharma 외 arxiv

Long-form fiction writers need memory that answers multi-hop questions about evolving story state: who knows a secret and when they learned it, whether an event preceded the narration that revealed it, whether a setup pa…

CraniMem: Cranial Inspired Gated and Bounded Memory for Agentic Systems

2026-03-03 · Pearl Mody, Mihir Panchal, Rishit Kar, Kiran Bhowmick 외 arxiv

Large language model (LLM) agents are increasingly deployed in long running workflows, where they must preserve user and task state across many turns. Many existing agent memory systems behave like external databases wit…

Membox: Weaving Topic Continuity into Long-Range Memory for LLM Agents

2026-01-07 · Dehao Tao, Guoliang Ma, Yongfeng Huang, Minghu Jiang arxiv

Long-term human-agent dialogues are organized by topic continuity: adjacent turns often develop the same goal, plan, problem, or event, while related activities may recur across distant sessions. Yet many LLM agent memor…

Semantic Retrieval

InfoMem: Training Long-Context Memory Agents with Answer-Conditioned Information Gain

2026-06-02 · Tiancheng Han, Yong Li, Wuzhou Yu, Qiaosheng Zhang 외 arxiv

Long-context tasks require LLMs to identify and preserve answer-relevant information from large contexts. Chunk-wise memory agents address this issue by sequentially reading document chunks, updating a compact memory, an…