[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82626-en":3,"doc-seo-82626-105":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},82626,16904993612988,"Olivia Brown","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","ADVENT: LLM-Driven Automatic Predicate Invention for ILP","Predicate invention (PI) is a major bottleneck in Inductive Logic Programming (ILP) because learning new relational rules depends on auxiliary predicates that expand the hypothesis space. Existing PI methods often require domain expertise and yield semantically opaque predicates, limiting adaptation to unfamiliar domains and reuse across tasks. ADVENT introduces an LLM-driven PI mechanism: it combines abductive predicate generation with Prolog deductive verification in an iterative loop where execution feedback refines candidates. Invented predicates and learned rules accumulate in a reusable knowledge pool. Experiments on poker-hand concepts across multiple LLMs report 58% success where ILP alone fails, 80% with formal verification, and up to +31 points from the pool, producing human-interpretable rules.","ADVENT: LLM-Driven Automatic Predicate  \nInvention for ILP  \nTingting Yu  \nDepartment of Information Management National Sun Yat-Sen University Kaohsiung, Taiwan[tingyui0213@gmail.com](tingyui0213@gmail.com)  \nPei-Cing Huang  \nDepartment of Information Management National Sun Yat-Sen University Kaohsiung, Taiwan [pcpeicing@gmail.com](pcpeicing@gmail.com)  \nChan Hsu  \nDepartment of Information Management National Sun Yat-Sen University Kaohsiung, Taiwan [chanshsu@gmail.com](chanshsu@gmail.com)  \nChan-Tung Ku  \nDepartment of Information Management National Sun Yat-Sen University Kaohsiung, Taiwan [kuchantung@gmail.com](kuchantung@gmail.com)  \nYihuang Kang  \nDepartment of Information Management National Sun Yat-Sen University Kaohsiung, Taiwan [ykang@mis.nsysu.edu.tw](ykang@mis.nsysu.edu.tw)  \narXiv :2607 .0 1585v 1 [ cs .LO] 2 Jul 2026  \nAbstract—Predicate invention (PI), the creation of new predicates to extend the hypothesis space, remains a critical bottleneck in Inductive Logic Programming (ILP). Existing methods rely on domain expertise and produce semantically opaque predicates, hindering adaptation to unfamiliar domains and cross-task reuse. We present ADVENT, an LLM-driven PI mechanism for ILP. ADVENT pairs LLM abductive generation with Prolog deductive verification, forming an iterative loop in which concrete execution results guide the LLM to refine candidate predicates. The mechanism leverages Large Language Models to identify implicit patterns in structured relational data and invent auxiliary predicates with meaningful names and definitions. Invented predicatesand learned rules accumulate in a knowledge pool for cross-task reuse. Experiments on nine poker-hand concepts across seven LLMs show that LLM-driven PI achieves 58% success rate where ILP alone fails entirely, formal verification raises this to 80%, and the knowledge pool yields gains up to +31 percentage points, while producing human-interpretable rules. These results suggest that ADVENT offers a promising direction for automating predicate invention and enabling cross-task knowledge reuse in ILP.  \nIndex Terms—Relational Rule Learning, Inductive Logic Programming, Predicate Invention, Large Language Model  \nI. INTRODUCTION  \nMany real-world concepts are inherently relational: a molecule is mutagenic not because of any single atom, but because of how atoms are bonded together. Inductive Logic Programming (ILP) [1] is one of the few machine learning paradigms capable of learning such relational concepts [2] as human-interpretable logical rules. However, ILP’s performance depends heavily on predefined background knowledge (BK) [1], a set of predicates that encode known facts and relational structures about the domain, analogous to features in traditional ML. Without suitable predicates, the system cannot discover rules that explain the target concept well. Predicate invention (PI) addresses this limitation by introducing auxiliary predicates that extend the hypothesis space beyond the original BK, making it one of the most critical open challenges in ILP [1] . Yet existing PI approaches suffer from two fundamental limitations. First, without human-provided specifications to guide the search, PI devolves into exhaustive syntactic enumeration over an explosive space, making it  \ndifficult to automatically discover helpful patterns [1], [3] . Second, existing PI methods operate through purely syntactic symbol manipulation and lack conceptual understanding of what they invent [4] . Unnamed predicates such as inv 1 and inv 2 have no notion of the patterns these symbols encode, making it hard to assess which prior inventions are relevant to a new task. Furthermore, predicates that build upon earlier inventions become increasingly difficult to interpret. Together, these limitations restrict ILP’s ability to learn in new domainsand hinder its potential for lifelong learning [5] .  \nLLMs offer a promising path to address these limitations. Trained on large-scale corpor","cbCaihdPHCs9EM74","https://ap.wps.com/l/cbCaihdPHCs9EM74","pdf",3373549,1,6,"English","en",105,"# Introduction\n## Motivation and Problem of Predicate Invention\n## Limitations of Existing Approaches\n## LLMs for Predicate Invention\n## Reliability via Deductive Verification\n## Research Questions and Proposed ADVENT","[{\"question\":\"What problem does ADVENT address in inductive logic programming?\",\"answer\":\"ADVENT targets predicate invention as a critical bottleneck in ILP. It aims to automatically create auxiliary predicates that extend the hypothesis space when predefined background knowledge is insufficient.\"},{\"question\":\"How does ADVENT generate and refine candidate predicates?\",\"answer\":\"ADVENT pairs LLM abductive generation with Prolog deductive verification. The system runs candidates on examples, uses execution results to guide iterative refinement, and improves generalization.\"},{\"question\":\"What benefits does ADVENT provide for cross-task reuse and interpretability?\",\"answer\":\"Invented predicates come with meaningful names and interpretable definitions, enabling relevance assessment across tasks. A knowledge pool accumulates invented predicates and learned rules to support reuse and composition while preserving human interpretability.\"}]",1784181894,15,{"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},"advent-llm-driven-automatic-predicate-invention-for-ilp","",{"@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/advent-llm-driven-automatic-predicate-invention-for-ilp/82626/",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},"What problem does ADVENT address in inductive logic programming?","Question",{"text":75,"@type":76},"ADVENT targets predicate invention as a critical bottleneck in ILP. It aims to automatically create auxiliary predicates that extend the hypothesis space when predefined background knowledge is insufficient.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does ADVENT generate and refine candidate predicates?",{"text":80,"@type":76},"ADVENT pairs LLM abductive generation with Prolog deductive verification. The system runs candidates on examples, uses execution results to guide iterative refinement, and improves generalization.",{"name":82,"@type":73,"acceptedAnswer":83},"What benefits does ADVENT provide for cross-task reuse and interpretability?",{"text":84,"@type":76},"Invented predicates come with meaningful names and interpretable definitions, enabling relevance assessment across tasks. A knowledge pool accumulates invented predicates and learned rules to support reuse and composition while preserving human interpretability.","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},[93,97,101,105,110,114,119,122,127,130,134],{"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":21,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"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"]