[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84624-en":3,"doc-seo-84624-105":29,"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":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84624,1649267921044,"Ava Thompson","https://us-avatar.wpscdn.com/avatar/1800007509477c92dfb?_k=1782875107921204101",8,"Research & Report","IntentTune Using User Demand and Personalization to Resolve Unknown Query Intents for E-Commerce Search","Understanding user intent is fundamental to delivering relevant search results in e-commerce, yet many real queries are underspecified, missing attributes like gender or age group. Such ambiguity challenges query intent detection, which must infer latent intent dimensions to support downstream retrieval and ranking. IntentTune resolves ambiguous query intents by conditioning on either user-specific behavioral signals (search history, browsing activity, profile attributes) or population-level demand patterns. Experiments show population signals alone are insufficient, while prior queries and behavioral signals outperform static profile information for key intent dimensions in fashion search.","IntentTune: Using user demand and personalization to resolve “unknown”  \nquery intents for e-commerce search  \nRachith Aiyappa, Ishita Khan, Chester Palen-Michel, Jayanth Yetukuri, Samarth Agrawal, Mehran Elyasi, Shuang Zhou  \neBay Inc., USA  \narXiv :2607 .0 1530v 1 [ cs .IR] 1 Jul 2026  \nAbstract  \nUnderstanding user intent is fundamental to delivering relevant search results in e-commerce. However, substantial fraction of real-world queries are under-specified (e.g.,“watch” or“shirt”), lacking explicit attributes such as gender or age group. This ambiguity poses a significant challenge for query intent detection models in e-commerce search systems, which must accurately infer latent user intent (e.g., age, gender) to support effective downstream retrieval. We introduce IntentTune, a framework for resolving ambiguous or underspecified query intents by leveraging either  \n(1) user-specific behavioral signals including search history, browsing activity, and profile attributes or (2) population-level demand patterns aggregated across all users. Through experiments on real-world e-commerce data, we first demonstrate that population-level demand patterns alone are insufficient to reliably infer intent in under-specified queries. We then demonstrate that user-specific behavioral signals—particularly prior search queries—outperform both population-level statistics and static profile information for inferring gender, age group, product category, and size intent from underspecified queries.  \n1 Introduction  \nRetrieving relevant products for a user query is a central problem in e-commerce search. Modern search engines typically rely on two complementary families of retrieval approaches:(i) keyword-based lexical retrieval methods such as inverted indexing and query-document lexical matching (Robertson et al., 1994), and (ii) embedding-based retrieval (EBR) methods that map queries and products into a shared semantic vector space (Lin et al., 2024 ; Huang et al., 2020, 2013 ; Kumar and Sarkar, 2021) . Despite their effectiveness, both paradigms struggle with the ambiguity inherent in under-specified user queries.  \nA substantial fraction of real-world queries are extremely short—often single tokens such as “boots,”“watch,” or “shirt” —and omit critical attributes such as gender, age group, style, or size. As a result, both lexical and semantic retrieval models frequently assign such queries to “unknown” or“unspecified” intent categories, or distribute them across coarse-grained buckets that fail to reflect the user’s true intent. This uncertainty propagates downstream, degrading recall, ranking quality, and overall user experience.  \nIn e-commerce settings, users exhibit strong and persistent preferences: browsing history, past purchases, saved items, and long-term category affinities provide powerful signals about what a user likely intends when issuing queries such as “boots”or “sneakers.” For instance, a query like “boots”from a user with a history of purchasing women’sankle boots conveys a markedly different intent than the same query issued by a user who typically shops for toddler footwear or men’s work boots. Ignoring such personalization signals leaves substantial intent information unexploited.  \nWe present IntentTune, a framework for resolving missing or ambiguous query intents produced by existing Query Understanding (QU) systems. IntentTune leverages two complementary sources of information: (1) population-level demand patterns, which capture aggregate trends across users, and (2) fine-grained user-specific signals, derived from individual browsing and interaction histories. This dual conditioning enables the inference of latent intent dimensions that are not expressed in the query text alone.  \nWe focus on fashion-related queries and show that IntentTune can reliably infer gender, age group, and size (when applicable) for a large fraction of previously unresolved queries. We first demonstrate that models based solely on pop","cbCaicGov2Xwyj0f","https://ap.wps.com/l/cbCaicGov2Xwyj0f","pdf",434828,1,9,"English","en",105,"# Introduction\n# Related Work\n## Query Understanding in E-Commerce\n## Personalization in Search and Recommendation","[{\"question\":\"Why do under-specified e-commerce queries lead to “unknown” intent categories?\",\"answer\":\"Many real queries are extremely short and omit crucial attributes such as gender, age group, style, or size. Lexical and embedding-based retrieval models then assign them to “unknown/unspecified” intent or coarse buckets that do not match the user’s true intent.\"},{\"question\":\"How does IntentTune resolve ambiguous query intents?\",\"answer\":\"IntentTune infers latent intent dimensions by conditioning on two complementary information sources: population-level demand patterns aggregated across users, and fine-grained user-specific behavioral signals from browsing and interaction histories.\"},{\"question\":\"Which signals work better for inferring gender, age group, product category, and size intent?\",\"answer\":\"Experiments show population-level demand patterns alone provide only modest accuracy. User-specific behavioral signals, especially prior search queries, consistently match or outperform both demand-based models and static profile information across multiple intent dimensions.\"}]",1784197232,23,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"intenttune-using-user-demand-and-personalization-to-resolve-unknown-query-intents-for-e-commerce-search","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/intenttune-using-user-demand-and-personalization-to-resolve-unknown-query-intents-for-e-commerce-search/84624/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","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},"Why do under-specified e-commerce queries lead to “unknown” intent categories?","Question",{"text":75,"@type":76},"Many real queries are extremely short and omit crucial attributes such as gender, age group, style, or size. Lexical and embedding-based retrieval models then assign them to “unknown/unspecified” intent or coarse buckets that do not match the user’s true intent.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does IntentTune resolve ambiguous query intents?",{"text":80,"@type":76},"IntentTune infers latent intent dimensions by conditioning on two complementary information sources: population-level demand patterns aggregated across users, and fine-grained user-specific behavioral signals from browsing and interaction histories.",{"name":82,"@type":73,"acceptedAnswer":83},"Which signals work better for inferring gender, age group, product category, and size intent?",{"text":84,"@type":76},"Experiments show population-level demand patterns alone provide only modest accuracy. User-specific behavioral signals, especially prior search queries, consistently match or outperform both demand-based models and static profile information across multiple intent dimensions.","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":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]