[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86136-en":3,"doc-seo-86136-105":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},86136,962075114765,"Quinn","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Generative Chinese Statute Retrieval","Statute retrieval is a core task in legal information retrieval, yet methods often fail to connect colloquial legal questions with formal statutory wording. This paper introduces GCSR, a generative statute retrieval framework that treats retrieval as sequence generation and embeds statutory knowledge into a generative model. A multi-granularity structured docid scheme captures legal hierarchy and semantics, paired with multitask training. Experiments show consistent gains over sparse, dense, and legal-domain baselines, supporting generative retrieval for broader legal access and reasoning.","Generative Chinese Statute Retrieval  \nYiteng Tu  \nTsinghua University Quancheng Laboratory Beijing, China [yitengtu16@gmail.com](yitengtu16@gmail.com)  \nYueyue Wu  \nTsinghua University Beijing, China  \nZitao Su  \nRenmin University of China Beijing, China  \nYiqun Liu  \nTsinghua University Beijing, China  \nWeihang Su  \nTsinghua University Beijing, China  \nMin Zhang  \nTsinghua University Beijing, China  \nXuanyi Chen  \nTsinghua University Beijing, China  \nQingyao Ai∗ Quancheng Laboratory Tsinghua University Beijing, China [aiqingyao@gmail.com](aiqingyao@gmail.com)  \narXiv :2607 . 1 1 109v 1 [ cs .IR] 13 Jul 2026  \nAbstract  \nStatute retrieval is a fundamental task in legal information retrieval, yet existing approaches struggle to bridge the gap between colloquial legal queries and formal statutory language. In this paper, we propose GCSR, a generative statute retrieval framework that reformulates statute retrieval as a sequence generation problem and internalizes statutory knowledge into a generative model. Specifically, we propose a multi-granularity structured docid that encodes legal hierarchy and semantic information, together with a multitask training strategy. Experiments show that GCSR consistently outperforms strong sparse, dense, and legal-domain baselines. Our results demonstrate the effectiveness of generative retrieval for statute retrieval and highlight its potential for broader legal information access and downstream legal reasoning tasks.  \nCCS Concepts  \n• Information systems → Retrieval models and ranking; • Applied computing → Law.  \nKeywords  \nStatute Retrieval, Generative Retrieval, Legal  \n1 Introduction  \nIn modern rule-of-law societies, the rapid growth of legal information has positioned Information Retrieval (IR) as a cornerstone of legal information systems [7, 11, 14, 17, 20] . From legal advisory services and due diligence to judicial decision support, practitioners and citizens routinely rely on search engines to access relevant legal materials [9, 19, 32] . While statutes are foundational across legal systems, many civil-law jurisdictions such as China place particular emphasis on codified provisions as the starting point for legal authority [2, 12] . Therefore, statute retrieval, the process of identifying relevant statutory articles for specific legal queries, serves as a core component of a wide range of downstream applications, including legal question answering, legal document drafting, and public legal consultation services [10, 22] . It is also essential to retrieval-augmented generation (RAG) [23, 24] systems for legal applications, where retrieved statutory provisions provide authoritative grounding for model-generated outputs. For legal professionals such as judges and lawyers, efficient retrieval systems  \n∗ Corresponding Author  \nsignificantly reduce the cognitive load of navigating vast legislative corpora. Moreover, for the general public, these systems provide a vital bridge to authoritative legal support, facilitating social justice and the accessibility of law.  \nHowever, accurately retrieving statutes remains challenging. On the one hand, everyday legal queries are typically formulated in colloquial, narrative language, whereas statutes are drafted using highly condensed and technical terminology [3, 18, 36]. Bridging this mismatch requires not only lexical matching but also the ability to interpret real-world narratives and map them to legally operative facts and applicable provisions. For example, a query stating that“my employer did not pay me for overtime work”must be connected to legally operative issues, such as the applicable working-hours regime and the employee’s entitlement to overtime compensation, even when the query contains no corresponding statutory terminology. On the other hand, statute retrieval necessitates complex legal reasoning that transcends simple keyword matching [10, 28], requiring an understanding of dependencies among provisions and the hierarchy of leg","cbCaimMdaX9FRgFq","https://ap.wps.com/l/cbCaimMdaX9FRgFq","pdf",498828,3,1,6,"English","en",105,"# 1 Introduction\n## Legal information retrieval context\n## Challenges: colloquial–statutory mismatch\n## Limitations of retrieve-then-rerank\n## Generative Retrieval and GCSR overview","[{\"question\":\"What problem does the paper address in statute retrieval?\",\"answer\":\"It targets the difficulty of bridging the semantic gap between colloquial legal queries and highly condensed, technical statutory language.\"},{\"question\":\"How does GCSR reformulate statute retrieval?\",\"answer\":\"GCSR models statute retrieval as an end-to-end sequence generation task that directly produces document identifiers conditioned on the query, using a generative retrieval framework.\"},{\"question\":\"What are the main components of the proposed approach?\",\"answer\":\"The method includes a multi-granularity structured docid scheme encoding legal hierarchy and semantics, plus a multitask training strategy with corpus training, pseudo-query search-oriented training, and supervised fine-tuning on annotated query–statute pairs.\"}]",1784208830,15,{"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},"generative-chinese-statute-retrieval","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"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":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/generative-chinese-statute-retrieval/86136/",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","2026-07-25","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 statute retrieval?","Question",{"text":75,"@type":76},"It targets the difficulty of bridging the semantic gap between colloquial legal queries and highly condensed, technical statutory language.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does GCSR reformulate statute retrieval?",{"text":80,"@type":76},"GCSR models statute retrieval as an end-to-end sequence generation task that directly produces document identifiers conditioned on the query, using a generative retrieval framework.",{"name":82,"@type":73,"acceptedAnswer":83},"What are the main components of the proposed approach?",{"text":84,"@type":76},"The method includes a multi-granularity structured docid scheme encoding legal hierarchy and semantics, plus a multitask training strategy with corpus training, pseudo-query search-oriented training, and supervised fine-tuning on annotated query–statute pairs.","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},[93,97,101,105,110,114,119,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"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":22,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"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":106,"slug":137},19,"General","general"]