[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82825-en":3,"doc-seo-82825-105":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},82825,5909877438554,"Maeve","https://ap-avatar.wpscdn.com/avatar/5600025385ad2bf12a7?_k=1778553567797529272",8,"Research & Report","From Interaction to Intent: Inferring User Objectives from Provenance Logs","Automatic inference of analytic intent from user interaction histories can enable proactive assistance during exploratory data analysis. This paper studies whether provenance logs—fine-grained records of interaction sequences and timing—can classify user intentions in visual exploration tasks. Participants perform multiple analytic objectives while interacting with multidimensional projections, capturing detailed mouse behavior across sessions. Distinct behavioral signatures emerge across objectives, and adding contextual information to provenance improves cross-dataset, cross-projection generalization, linking low-level actions to high-level intent for intent-aware visualization.","From Interaction to Intent: Inferring User Objectives from Provenance Logs  \nSteffen Holter  \n[steffen.holter@inf.ethz.ch](steffen.holter@inf.ethz.ch)[ ](steffen.holter@inf.ethz.ch)ETH Zurich Zurich, Switzerland  \nTobias Stähle  \n[tobias.staehle@inf.ethz.ch](tobias.staehle@inf.ethz.ch)[ ](tobias.staehle@inf.ethz.ch)ETH Zurich Zurich, Switzerland  \narXiv :2607 .0450 1v 1 [ cs .HC] 5 Jul 2026  \nArpit Narechania  \n[arpit@ust.hk](arpit@ust.hk)  \nThe Hong Kong University of Science and Technology Hong Kong S.A.R., China  \nMennatallah El-Assady  \n[melassady@ai.ethz.ch](melassady@ai.ethz.ch)[ ](melassady@ai.ethz.ch)ETH Zurich Zurich, Switzerland  \nFigure 1: Summary of the Interactions-to-Intent approach: A three-stage pipeline for analyzing and predicting user interaction patterns during multidimensional projection exploration tasks. (1) 1439 usable interaction sequences are crowdsourced for atomic exploration tasks. (2) Raw interaction logs are contextualized using either a summary or temporal representation, and modeled for both atomic task classification and online prediction of multi-task sessions. (3) The resulting models are used to infer user objectives (i.e., tasks) and analyze behavioral patterns across tasks.  \nAbstract  \nThe ability to automatically infer analytic intent from user interaction histories could enable interactive AI systems to proactively assist users during exploratory data analysis. In this paper, we examine whether provenance logs – detailed records capturing sequencesand timing of user interactions – can be used to classify user intentions in visual exploration tasks. To investigate this, we record how participants interact with multiple multidimensional data projections across a range of analytic tasks, capturing fine-grained mouse interaction data throughout each session. We find that distinct behavioral signatures emerge across different analytic objectives. For instance, users examining properties of specific clusters exhibit markedly different interaction patterns compared to those searching for outliers. More importantly, we show that embedding contextual information into interaction provenance enables classifiersto predict user objectives that generalize across datasets and projection methods. These findings demonstrate that low-level interaction data can serve as a practical bridge to high-level analytic intent, contributing to the development of intent-aware visualization systems.  \nCCS Concepts  \n• Human-centered computing → Visualization.  \nKeywords  \nvisualization, intent inference, interaction, machine learning  \n1 Introduction  \nIn visual analytics (VA), most interactive AI tools remain predominantly reactive, responding explicitly to user inputs rather than proactively supporting analytical workflows. However, if a system could implicitly infer user objectives from their actions, it could offer more dynamic and intent-aligned support. The challenge lies in accurately interpreting user goals and thought processes without introducing additional interaction overhead. Fortunately, when people use interactive systems, they inevitably generate a trail of interaction data that can provide insight into their cognitive processes and decision-making. Provenance logs – detailed records capturing the sequence, timing, and nature of user interactions – have traditionally been used to deconstruct analytical behavior post hoc and to gain insights into user reasoning processes. However, thus far, little work has examined whether these interaction patterns can be leveraged to predict user intentions (i.e., what they are attempting to accomplish) and behavior across a large set of users.  \nIn this paper, we investigate how histories of user interactions can inform predictions about user objectives during visual exploration tasks. By crowdsourcing provenance logs of mouse activity at scale, we construct a dataset that captures how users interact with scatter plots during tasks. In our study, we focus on a wel","cbCaih0bcs35RpJx","https://ap.wps.com/l/cbCaih0bcs35RpJx","pdf",9056443,5,1,19,"English","en",105,"# Introduction\n## Motivation and challenge in visual analytics\n## Provenance logs for intent prediction\n## Crowdsourced dataset and task scenario\n## Contextualization for cross-layout generalization","[{\"question\":\"What problem does the paper address in visual analytics?\",\"answer\":\"Most interactive AI tools are primarily reactive. The paper addresses how to implicitly infer user objectives from interaction histories so systems can provide more dynamic, intent-aligned support.\"},{\"question\":\"How are provenance logs used in the study?\",\"answer\":\"Provenance logs capture detailed sequences and timing of user interactions, including fine-grained mouse behavior during exploratory visual tasks. These logs are used to classify user intentions.\"},{\"question\":\"What is the main finding about behavioral signatures and contextualization?\",\"answer\":\"Users exhibit distinct interaction signatures for different analytic objectives (e.g., examining clusters versus searching for outliers). Contextualizing provenance with features enables classifiers to generalize across datasets and projection methods.\"}]",1784183216,48,{"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},"from-interaction-to-intent-inferring-user-objectives-from-provenance-logs","",{"@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/from-interaction-to-intent-inferring-user-objectives-from-provenance-logs/82825/",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},"What problem does the paper address in visual analytics?","Question",{"text":76,"@type":77},"Most interactive AI tools are primarily reactive. The paper addresses how to implicitly infer user objectives from interaction histories so systems can provide more dynamic, intent-aligned support.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How are provenance logs used in the study?",{"text":81,"@type":77},"Provenance logs capture detailed sequences and timing of user interactions, including fine-grained mouse behavior during exploratory visual tasks. These logs are used to classify user intentions.",{"name":83,"@type":74,"acceptedAnswer":84},"What is the main finding about behavioral signatures and contextualization?",{"text":85,"@type":77},"Users exhibit distinct interaction signatures for different analytic objectives (e.g., examining clusters versus searching for outliers). Contextualizing provenance with features enables classifiers to generalize across datasets and projection methods.","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},[94,98,102,106,110,115,120,123,128,131,135],{"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":20,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},"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":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":20,"slug":137},"General","general"]