[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82190-en":3,"doc-seo-82190-105":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},82190,2336464648746,"Skyler","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","PRecG: Legal Precedent Retrieval with Graph Neural Networks and Rhetorical Role Segmentation","Legal precedent retrieval underpins legal case preparation, litigation planning, and research by matching judgments based on semantic similarity. Existing methods embed entire legal documents into low-dimensional spaces, treating them as monolithic texts and overlooking rhetorical organization, which obscures nuanced meanings and entity significance across contexts. PRecG addresses this by segmenting documents using rhetorical roles, building knowledge graphs per segment, learning entity-context representations, aggregating into unified document embeddings, and computing similarity. Experiments on an Indian legal benchmark validate performance versus state-of-the-art baselines.","arXiv :2607 .09094v 1 [ cs .CL] 10 Jul 2026  \nPRecG: Legal Precedent Retrieval with Graph Neural Networksand Rhetorical Role Segmentation  \nDevanshu Vermaa , Vasudha Bhatnagara , Vikas Kumara,∗, Balaji Ganesanb  \na University of Delhi, Delhi, India  \nb IBM Research, Bengaluru, India  \nAbstract  \nLegal precedent retrieval is a fundamental task in legal case preparation, planning, litigation strategy, and legal research. Current approaches for automatic precedent retrieval map legal documents to a low-dimensional semantic space and compute similarity based on the proximity of their representations. These approaches treat legal documents as monolithic texts, ignoring the rhetorical organization of the legal technicalities. Ergo, they overlook nuanced legal meanings and fail to distinguish the contextual significance of legal entities and concepts that vary based on their rhetorical roles within the document.  \nTo address this insufficiency, we propose the PRecG pipeline that computes the similarity between pairs of legal judgments by hierarchically learning their representations. The process begins by decomposing each document into distinct semantic units (segments) based on the rhetorical roles of sentences. For each rhetorical segment, a knowledge graph is constructed to capture the legal entities and their relationships within the segment. Contextual representations of the entities are then learned and aggregated to derive segment-level embeddings. These embeddings are further integrated to produce a unified document-level representation, and finally, the semantic similarity between a pair of documents is computed. We validate the performance of the proposed approach through extensive experiments on a benchmark Indian legal dataset, comparing it against state-of-the-art baselines to demonstrate its effectiveness.  \nKeywords: Legal Text Analytics, Precedence Retrieval, Similar Case Retrieval, Rhetorical Role  \n1. Introduction  \nLegal systems aim to ensure that human actions within society are kept in order and control. They form an integral part of a nation’s culture, civilization, history, and the everyday life of its people. Different countries around the world have various jurisdictions,  \n∗ Corresponding author  \nEmail addresses: [dverma@cs.du.ac.in](dverma@cs.du.ac.in) (Devanshu Verma), [vbhatnagar@cs.du.ac.in](vbhatnagar@cs.du.ac.in) (Vasudha  \nBhatnagar), [vikas@cs.du.ac.in](vikas@cs.du.ac.in) (Vikas Kumar), [bganesa1@in.ibm.com](bganesa1@in.ibm.com) (Balaji Ganesan)  \nFigure 1: Manual procedure for finding valid and reliable citations.  \nincluding civil law, common law, customary law, religious law, and mixed law. India, in particular, operates under a mixed legal system, with the common law having a prominent role. This system adheres to the concept of stare decisis, which accords equal importance to codified statutes and prior case judgments [1] . The fundamental principle is that cases with similar facts and situations should be decided consistently, in accordance with the judgments of prior similar cases or precedents.  \nPrecedent cases are integral to the common law system and serve as the basis for judicial decision-making [2] . Drawing upon prior judgments to deliver verdicts promotes consistency, transparency, and fairness in legal proceedings [1] . Therefore, ensuring the relevance and reliability of precedents is essential and must be strongly adhered. Typically, the process of precedent selection begins with a keyword-based search from an e-repository of legal cases, and a trained professional manually selects a set of relevant keywords as a query to retrieve a candidate set of cases. The search engine retrieves potentially relevant documents from the repository, and the legal experts carefully review and rank the retrieved cases. Often, a consultation process results in the final selection of precedent cases. This process, depicted in Figure 1, is inherently time-consuming and prone to limitations arising fr","cbCaibP9bGKmzAHj","https://ap.wps.com/l/cbCaibP9bGKmzAHj","pdf",937266,1,23,"English","en",105,"# Introduction\n## Motivation for automatic precedent retrieval\n## Challenges in current embedding-based methods\n# Proposed PRecG approach","[{\"question\":\"Why do existing automatic precedent retrieval methods struggle with legal texts?\",\"answer\":\"They typically treat legal documents as monolithic texts, so they miss how rhetorical organization affects meaning and how entity significance changes across contexts and sections.\"},{\"question\":\"What is the core idea behind the PRecG pipeline?\",\"answer\":\"PRecG computes similarity between legal judgments by hierarchically learning representations, starting with rhetorical-role-based segmentation and then building knowledge graphs for each segment.\"},{\"question\":\"How does PRecG derive document-level similarity from segment-level information?\",\"answer\":\"For each rhetorical segment, it learns contextual entity representations and aggregates them into segment embeddings, integrates these to form unified document embeddings, and finally computes similarity between document pairs.\"}]",1784178703,58,{"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},"precg-legal-precedent-retrieval-with-graph-neural-networks-and-rhetorical-role-segmentation","",{"@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/precg-legal-precedent-retrieval-with-graph-neural-networks-and-rhetorical-role-segmentation/82190/",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 existing automatic precedent retrieval methods struggle with legal texts?","Question",{"text":75,"@type":76},"They typically treat legal documents as monolithic texts, so they miss how rhetorical organization affects meaning and how entity significance changes across contexts and sections.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is the core idea behind the PRecG pipeline?",{"text":80,"@type":76},"PRecG computes similarity between legal judgments by hierarchically learning representations, starting with rhetorical-role-based segmentation and then building knowledge graphs for each segment.",{"name":82,"@type":73,"acceptedAnswer":83},"How does PRecG derive document-level similarity from segment-level information?",{"text":84,"@type":76},"For each rhetorical segment, it learns contextual entity representations and aggregates them into segment embeddings, integrates these to form unified document embeddings, and finally computes similarity between document 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":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"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":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]