[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81712-en":3,"doc-seo-81712-105":30,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},81712,3848291630094,"Emma Wilson","https://eur-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","Learning User-Aware Recall: Personalized Retrieval in Long-Term Conversational Memory","Long-term conversational agents must recall useful past interactions only when the evidence matches the right user. Prior memory-augmented LLM systems often retrieve via query similarity or fixed ranking, leaving user-conditioned relevance insufficiently explored. This work introduces Profile-guided Personalized Retrieval Optimization (PPRO), which builds episodic and semantic memory banks and derives a user profile from accumulated memories. The profile acts as an explicit personalized prior in memory ranking, while a GRPO-trained query rewriter uses retrieval quality and downstream answer quality feedback. Experiments on LoCoMo and LongMemEval-S show consistent gains, with ablations confirming retrieval-oriented optimization as a key driver.","Learning User-Aware Recall:  \nPersonalized Retrieval in Long-Term Conversational Memory  \nZhiShu Jiang 1 , Haibo Liu 1 , Xin Shen 1,2 , Guanqiang Qi 1 , Chenxi Miao 1 ,  \nWeikang Li 1,†, Liwei Qian 1 , Xin Pei 1 , Jizhou Huang 1  \n1Baidu Inc. 2The University of Queensland  \n[jiangzhishu@bjtu.edu.cn](jiangzhishu@bjtu.edu.cn) , [wavejkd@pku.edu.cn](wavejkd@pku.edu.cn) *  \narXiv :2607 .000 17v2 [ cs .IR] 2 Jul 2026  \nAbstract  \nLong-term conversational agents are expected to remember past interactions, but memory is useful only when the right evidence is recalled for the right user. Existing memory-augmented LLM agents have made progress in building compact memory banks, yet retrieval is still often driven by query-centered similarity or fixed ranking rules, leaving user-conditioned relevance underexplored. To address this gap, we propose Profile-guided Personalized Retrieval Optimization (PPRO), a retrievalcentric framework that makes memory retrieval both user-aware and optimizable. PPRO builds episodic and semantic memory banks from dialogue histories and derives a user profile from accumulated memories. The profile serves as an explicit personalized prior in memory ranking, allowing retrieval to account for stable user attributes, preferences, and relationships. PPRO further trains a query rewriter with Group Relative Policy Optimization, using both evidence retrieval quality and downstream answer quality as feedback while keeping the memory banks and answer model fixed. Experiments on LoCoMo and LongMemEval-S show consistent gains over training-free memory systems and training-based baselines. Ablation studies further show that both profileguided ranking and retrieval-oriented rewriting contribute substantially to performance, highlighting retrieval optimization as a key factor in personalized long-term memory use.  \n1 Introduction  \nLarge language model (LLM) agents are increasingly expected to support long-term personalized interaction, where useful knowledge must be accumulated across extended dialogue histories and reused in later conversations (Maharana et al., 2024 ; Packer et al., 2023 ; Wu et al., 2026) . Recent  \n*† Corresponding author  \nFigure 1: Motivating example of profile-guided personalized memory retrieval. Query-only retrieval may return surface-matching evidence, while profile-guided retrieval uses the user’s preferences, plans, and relationships to produce a more personalized answer.  \nmemory-augmented systems have made substantial progress in reducing the cost of long-context inputs by extracting, compressing, and organizing historical interactions into external memory banks (Chhikara et al., 2025 ; Xu et al., 2025 ; Liu et al., 2026 ; Kang et al., 2025 ; Lewis et al., 2020 ; Shen et al., 2026 ; Liao et al., 2026) . As these systems mature, a central challenge moves from memory construction alone to memory use at inference time: given a user query, the agent must retrieve evidence that is not only semantically relevant to the  \nquery, but also appropriate for the user behind the query (Shen et al., 2021c) . In long-term dialogue, different users accumulate distinct memories that reflect their relationships, topical interests, and life contexts (Shen et al., 2021a,b) . The same query may therefore correspond to different information needs depending on who is asking. As illustrated in Figure 1, personalized memory retrieval requires going beyond surface query matching: the retriever must identify evidence that is jointly relevant to the query and to the user’s long-term profile.  \nExisting memory systems have not fully addressed this user-conditioned retrieval problem. Most work focuses on memory construction and management, including extraction, compression, hierarchical organization, graph-based indexing, and read-write operations (Liu et al., 2026 ; Xu et al., 2025 ; Chhikara et al., 2025) . These methods improve the coverage and efficiency of stored memories, but retrieval is often treated as a fixed d","cbCaim2SlHIOHOWZ","https://ap.wps.com/l/cbCaim2SlHIOHOWZ","pdf",961561,6,1,16,"English","en",105,"# Abstract\n# Introduction\n## Problem: user-conditioned evidence retrieval\n## Limitations of existing memory systems\n## Proposed approach: PPRO","[{\"question\":\"Why is user-aware recall important in long-term conversational agents?\",\"answer\":\"Memory helps only when the agent retrieves the right evidence for the right user. Different users accumulate distinct memories, so the same query can imply different information needs.\"},{\"question\":\"How does PPRO make retrieval personalized?\",\"answer\":\"PPRO constructs episodic and semantic memory banks and derives a user profile from accumulated memories. It injects the profile as an embedding-level personalized prior into retrieval ranking.\"},{\"question\":\"How is personalization optimized during training in PPRO?\",\"answer\":\"PPRO trains a query rewriter using Group Relative Policy Optimization (GRPO), leveraging feedback from both retrieval evidence quality and downstream answer quality while keeping the memory banks and answer model fixed.\"}]",1784175569,40,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"learning-user-aware-recall-personalized-retrieval-in-long-term-conversational-memory","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/learning-user-aware-recall-personalized-retrieval-in-long-term-conversational-memory/81712/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"Why is user-aware recall important in long-term conversational agents?","Question",{"text":76,"@type":77},"Memory helps only when the agent retrieves the right evidence for the right user. Different users accumulate distinct memories, so the same query can imply different information needs.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does PPRO make retrieval personalized?",{"text":81,"@type":77},"PPRO constructs episodic and semantic memory banks and derives a user profile from accumulated memories. It injects the profile as an embedding-level personalized prior into retrieval ranking.",{"name":83,"@type":74,"acceptedAnswer":84},"How is personalization optimized during training in PPRO?",{"text":85,"@type":77},"PPRO trains a query rewriter using Group Relative Policy Optimization (GRPO), leveraging feedback from both retrieval evidence quality and downstream answer quality while keeping the memory banks and answer model fixed.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,115,119,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":29,"slug":118},7,"Healthcare","healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":107,"slug":137},19,"General","general"]