[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83511-en":3,"doc-seo-83511-105":29,"detail-sidebar-cat-0-en-105":83},{"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},83511,549758146520,"Patrick","https://ap-avatar.wpscdn.com/avatar/80002397d8c0411e94?_k=1775819394049821470",8,"Research & Report","Multi-Turn Agentic Scientific Literature Search via Workflow Induction","Scientific literature search often requires more than retrieving papers from a single query: user intents are underspecified, evolve through interaction, and depend on nuanced preferences. Existing search agents use fixed pipelines or implicit reasoning, limiting control, inspection, and refinement. PAPERPILOT frames literature search as workflow induction: from an anchor paper and a query, it builds an executable DAG of operators covering search, expansion, filtering, scoring, reranking, and evidence extraction. Feedback refines both query and workflow, improving Hit@5, MRR, nDCG@10 while eliminating workflow execution errors.","Multi-Turn Agentic Scientific Literature Search via Workflow Induction  \nJisen Li1,2 *† Bingxuan Li1 * Xuying Ning 1 Xiyao Wang3 Yifan Shen 1 Xiaoxia Wu2 Ben Athiwaratkun2 Pan Lu4  \nNanyi Jiang3 * Heng Wang 1 Jiaxuan You 1  \nYuqing Jian2 Bingxin Zhao3  \n1University of Illinois Urbana-Champaign 2Together AI  \n3University of Pennsylvania 4 Stanford University  \n􀂀 Project Website § Code  \narXiv :2607 .00597v2 [ cs .CL] 3 Jul 2026  \nAbstract  \nScientific literature search often requires more than retrieving papers from a single query:  \nusers’ intents are underspecified, preferencedependent, and evolve through interaction. Existing search agents typically rely on fixed pipelines or implicit language-only reasoning, making their search strategies difficult to control, inspect, and refine. We introduce PAPERPILOT, a multi-turn literature search agent that frames scientific search as workflow induction. Given an anchor paper and a user query, PAPERPILOT constructs an executable DAG of papersearch operators, including keyword search, citation expansion, filtering, scoring, reranking, and evidence extraction. User feedback is then used to refine both the query and the workflow itself. We train PAPERPILOT with supervised workflow imitation and preference optimization over controlled workflow corruptions. Experiments show that PAPERPILOT-9B improves over the base Qwen3.5-9B toolset agent under multi-turn interaction, increasing Hit@5 from 58.0 to 77.0, MRR from 47.5 to 59.4, and nDCG@10 from 26.8 to 32.5, while reducing workflow execution errors from 9 .5% to 0% . These results show that explicit, editable search workflows provide an effective and controllable interface for aligning literature search agents with complex scientific intent.  \n1 Introduction  \nScientific literature search is a core research activity that increasingly requires more than keyword matching. Researchers must identify relevant prior work, explore citation and semantic neighborhoods, compare candidate papers, filter distractors, and understand why retrieved papers are relevant to a given research intent. As scientific output continues to grow, literature search increasingly depends  \n*Equal contribution, order interchangeable.†Contact: {jisenli2, [bl61}@illinois.edu](bl61}@illinois.edu)  \n(a) Single turn scientific literature search  \nunrelated result costly one-shot search  \n(b) Multi-turn scientific literature search  \nrelated result cost-effective workflow  \nFigure 1: Overview of PAPERPILOT. Compared with single-turn scientific literature search, PAPERPILOT uses multi-turn feedback to clarify user intent and refine the search workflow before producing final results.  \non adaptive retrieval and interaction rather than static query matching.  \nRecent agentic search systems improve retrieval quality by combining language models with external tools, iterative reasoning, and multi-step search (Agarwal et al., 2024 ; Skarlinski et al., 2024 ; Asai et al., 2024 ; He et al., 2025 ; Baek et al., 2025) . However, search intent is rarely fully specified in a single query, echoing broader findingsin clarification-driven retrieval that ambiguous information needs often require interactive disambiguation (Zamani et al., 2020 ; Li et al., 2025a ; Chi et al., 2024 ; Liu et al., 2026a ; Li et al., 2026b ; Wang et al., 2026) . For example, a request such as “find follow-up work on this paper” may refer to direct citations, recent extensions, papers in the same application domain, or work building upon a specific methodological component. The correct retrieval strategy therefore depends on latent user preferences and evolving interaction feedback.  \n\n| System | Symbolic\u003Cbr>Workflow | Workflow\u003Cbr>Refinement | Multi-turn\u003Cbr>Dialogue | Citation\u003Cbr>Expansion | Evidence\u003Cbr>Grounding |\n| --- | --- | --- | --- | --- | --- |\n| OpenAI DeepResearch (OpenAI, 2025) | × | × | △ | − | △ |\n| LitLLM (Agarwal et al., 2024) | × | × | × | × | × |\n| STORM (Shao et al., 2024) | × | × | × | × | △ |\n| ","cbCaimTHphkB4g8b","https://ap.wps.com/l/cbCaimTHphkB4g8b","pdf",2111649,1,17,"English","en",105,"# Abstract\n# Introduction\n## Single-turn vs multi-turn literature search\n## Motivation and limitations of existing agents\n## Proposed approach: PAPERPILOT","[{\"question\":\"How is PAPERPILOT trained and what improvements does it achieve?\",\"answer\":\"It is trained with supervised workflow imitation and preference optimization over controlled workflow corruptions. Experiments show higher Hit@5, MRR, and nDCG@10 compared with a base toolset agent, with fewer workflow execution errors.\"}]",1784188537,43,{"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":78,"head_meta":80,"extra_data":82,"updated_unix":27},"multi-turn-agentic-scientific-literature-search-via-workflow-induction","",{"@graph":35,"@context":77},[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/multi-turn-agentic-scientific-literature-search-via-workflow-induction/83511/",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],{"name":72,"@type":73,"acceptedAnswer":74},"How is PAPERPILOT trained and what improvements does it achieve?","Question",{"text":75,"@type":76},"It is trained with supervised workflow imitation and preference optimization over controlled workflow corruptions. Experiments show higher Hit@5, MRR, and nDCG@10 compared with a base toolset agent, with fewer workflow execution errors.","Answer","https://schema.org",{"og:url":51,"og:type":79,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":81,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":84},[85,89,93,97,102,107,112,115,120,123,127],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":86,"show_sort_weight":87,"slug":88},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":90,"show_sort_weight":91,"slug":92},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Exam",70,"exam",{"id":98,"doc_module":4,"doc_module_name":45,"category_name":99,"show_sort_weight":100,"slug":101},5,"Comic",60,"comic",{"id":103,"doc_module":4,"doc_module_name":45,"category_name":104,"show_sort_weight":105,"slug":106},6,"Technology",50,"technology",{"id":108,"doc_module":4,"doc_module_name":45,"category_name":109,"show_sort_weight":110,"slug":111},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":113,"slug":114},30,"research-report",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":45,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":128,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":98,"slug":130},19,"General","general"]