{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/incremental-learning-of-event-definitions","title":"Incremental Learning of Event Definitions with Inductive Logic Programming","arxiv_id":"1402.5988","date":"2014-02-24","proceeding":null,"authors":["Nikos Katzouris","Alexander Artikis","George Paliouras"],"abstract":"Event recognition systems rely on properly engineered knowledge bases of\nevent definitions to infer occurrences of events in time. The manual\ndevelopment of such knowledge is a tedious and error-prone task, thus\nevent-based applications may benefit from automated knowledge construction\ntechniques, such as Inductive Logic Programming (ILP), which combines machine\nlearning with the declarative and formal semantics of First-Order Logic.\nHowever, learning temporal logical formalisms, which are typically utilized by\nlogic-based Event Recognition systems is a challenging task, which most ILP\nsystems cannot fully undertake. In addition, event-based data is usually\nmassive and collected at different times and under various circumstances.\nIdeally, systems that learn from temporal data should be able to operate in an\nincremental mode, that is, revise prior constructed knowledge in the face of\nnew evidence. Most ILP systems are batch learners, in the sense that in order\nto account for new evidence they have no alternative but to forget past\nknowledge and learn from scratch. Given the increased inherent complexity of\nILP and the volumes of real-life temporal data, this results to algorithms that\nscale poorly. In this work we present an incremental method for learning and\nrevising event-based knowledge, in the form of Event Calculus programs. The\nproposed algorithm relies on abductive-inductive learning and comprises a\nscalable clause refinement methodology, based on a compressive summarization of\nclause coverage in a stream of examples. We present an empirical evaluation of\nour approach on real and synthetic data from activity recognition and city\ntransport applications.","url_abs":"http://arxiv.org/abs/1402.5988v2","url_pdf":"http://arxiv.org/pdf/1402.5988v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"incremental-learning-of-event-definitions","repo_url":"https://github.com/nkatzz/iled","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"activity-recognition","task_name":"Activity Recognition"},{"task_slug":"incremental-learning","task_name":"Incremental Learning"},{"task_slug":"inductive-learning","task_name":"Inductive Learning"},{"task_slug":"inductive-logic-programming","task_name":"Inductive logic programming"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}