[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83140-en":3,"doc-seo-83140-105":30,"detail-sidebar-cat-0-en-105":91},{"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},83140,687197207057,"Sage","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","When and How to Ask Dynamic Preference Elicitation Strategies for Conversational Recommendation","Conversational Recommender Systems use multi-turn dialogue to track evolving user preferences and provide personalized suggestions, yet the timing and choice of preference elicitation strategies remain underexplored. This work studies strategy selection from a stage-aware perspective, showing that attribute-based questions excel early while item-based strategies outperform as preferences become more refined. It introduces InPE, a dataset with fine-grained annotations for elicitation necessity and strategy choice, and proposes COPE, a mixture-of-experts modeling architecture, validated by extensive offline experiments.","When and How to Ask: Dynamic Preference Elicitation Strategies  \nfor Conversational Recommendation  \nFeng Xia  \nUniversity of Sheffield Sheffield, United Kingdom [fxia8@sheffield.ac.uk](fxia8@sheffield.ac.uk)  \nShuo Zhang  \nBloomberg London, United Kingdom[szhang611@bloomberg.net](szhang611@bloomberg.net)  \nXi Wang  \nUniversity of Sheffield Sheffield, United Kingdom [xi.wang@sheffield.ac.uk](xi.wang@sheffield.ac.uk)  \narXiv :2607 .06765v 1 [ cs .IR] 7 Jul 2026  \nAbstract  \nConversational Recommender Systems (CRSs) are interactive systems that use multi-turn natural language dialogue to understand evolving user preferences and provide personalized recommendations. To achieve this goal, CRSs rely on preference elicitation strategies to actively gather informative preference cues from users; however, the timing and selection of these strategies during a conversation remain largely unexplored. While many existing studies emphasize eliciting explicit item attributes and tend to adopt relatively static elicitation strategies, the use of item-based preference elicitation and how it varies across different dialogue stages remains less explored. In this work, we conduct a systematic investigation of preference elicitation strategies from a stage-aware perspective. We provide empirical evidence that optimal preference elicitation strategies are stage-dependent and context-sensitive: attribute-based inquiries are effective in early stages, while itembased strategies become superior as preferences refine. To support this paradigm, we introduce InPE, a dataset enriched with finegrained annotations for elicitation necessity and strategy selection. With this dataset, we propose COPE (COnversational Preference Elicitation via Mixture of Experts), a novel architecture for strategy modeling. Extensive offline evaluation on our dataset indicates that context-aware preference elicitation strategies are beneficial for conversational recommendation. In addition, the analysis of the predicted strategies uncovers consistent stage-wise tendencies in dialogue progression, providing empirical evidence of common interaction patterns in conversational recommendation systems. Our dataset is available at [https://github.com/juanfacabian/InPE](https://github.com/juanfacabian/InPE).  \nCCS Concepts  \n• Information systems → Personalization.  \nKeywords  \nConversational Recommender Systems, Preference Elicitation, Proactive Conversational Systems  \n1 Introduction  \nConversational Recommender Systems (CRSs) deliver personalized recommendations through natural language interactions with users. A core advantage of CRSs lies in their ability to actively collect user preferences during a conversation, enabling more accurate and adaptive personalized recommendations. However, capturing users’ evolving preferences and determining when and how to elicit them in a conversation remain fundamental challenges in building effective CRSs.  \nTraditional recommender systems infer user preferences primarily from historical user behavior [22], sometimes augmented  \nUser Cho ice Ratio  \n0.70  \n0.65  \n0.60  \n0.55  \n0.50  \n0.45  \n0.40  \n0.35  \n0.30  \n Attribute-based  Item-based  \n1 2 3 4 5 6 7 8 Preference Evolution (from Abstract to Concrete)  \nFigure 1: User selection ratios of preference elicitation strategies across dialogue stages. The x-axis shows the progression from abstract to concrete preferences (Stages 1–8), and they-axis denotes the proportion of attribute-based and itembased strategy selections.  \nwith user-centric knowledge graphs [26]. While effective in many scenarios, these approaches are significantly limited by noisy and sparse collaborative signals. For example, users may click on items they do not actually like. In contrast, CRSs exhibit greater proactiveness and flexibility in collecting user preferences rather than relying solely on human feedback. One fundamental way to collect user preferences in CRSs is through preference elicitation methods. Unlike appro","cbCailT7n9ud29Lf","https://ap.wps.com/l/cbCailT7n9ud29Lf","pdf",986912,2,1,10,"English","en",105,"# Abstract\n# Introduction\n## Problem motivation\n## Preference elicitation methods\n## Attribute-based vs. item-based approaches\n# Dynamic preference elicitation across dialogue stages\n## Stage-wise strategy tendencies\n# Method and contributions\n## InPE dataset\n## COPE model (mixture of experts)\n# Evaluation and results","[{\"question\":\"What problem does the paper address in conversational recommender systems?\",\"answer\":\"It addresses when and how to elicit user preferences during multi-turn dialogue, since current systems often use a single elicitation strategy throughout the conversation.\"},{\"question\":\"How do attribute-based and item-based preference elicitation strategies differ across dialogue stages?\",\"answer\":\"Attribute-based inquiries are effective in early stages, while item-based strategies become superior as user preferences are refined later in the dialogue.\"},{\"question\":\"What are InPE and COPE, and what role do they play?\",\"answer\":\"InPE is a dataset with fine-grained annotations for elicitation necessity and strategy selection. COPE is a proposed mixture-of-experts architecture for modeling and selecting elicitation strategies, supported by extensive offline evaluation.\"}]",1784185561,25,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":28},"when-and-how-to-ask-dynamic-preference-elicitation-strategies-for-conversational-recommendation","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,47,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":20},"https://docshare.wps.com/document/","Document",{"item":48,"name":12,"@type":43,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/when-and-how-to-ask-dynamic-preference-elicitation-strategies-for-conversational-recommendation/83140/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-22","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address in conversational recommender systems?","Question",{"text":75,"@type":76},"It addresses when and how to elicit user preferences during multi-turn dialogue, since current systems often use a single elicitation strategy throughout the conversation.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How do attribute-based and item-based preference elicitation strategies differ across dialogue stages?",{"text":80,"@type":76},"Attribute-based inquiries are effective in early stages, while item-based strategies become superior as user preferences are refined later in the dialogue.",{"name":82,"@type":73,"acceptedAnswer":83},"What are InPE and COPE, and what role do they play?",{"text":84,"@type":76},"InPE is a dataset with fine-grained annotations for elicitation necessity and strategy selection. COPE is a proposed mixture-of-experts architecture for modeling and selecting elicitation strategies, supported by extensive offline evaluation.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":22,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]