{"type":"ScholarlyArticle","title":"Practical Partition-Based Theorem Proving for Large Knowledge Bases","authors":[{"name":"Bill MacCartney","slug":"bill-maccartney","given":"Bill","family":"MacCartney","affiliation":""},{"name":"Sheila A. McIlraith","slug":"sheila-a-mcilraith","given":"Sheila A.","family":"McIlraith","affiliation":""},{"name":"Eyal Amir","slug":"eyal-amir","given":"Eyal","family":"Amir","affiliation":""},{"name":"Tomás E. Uribe","slug":"tomas-e-uribe","given":"Tomás E.","family":"Uribe","affiliation":""}],"abstract":"Query answering over commonsense knowledge bases typically employs a first-order logic theorem prover. While first-order inference is intractable in genera l, provers can often be hand-tuned to answer queries with reasonable performance in practice. Appealing to previous theoretical work on partition-based reasoning, we propose resolutionbased theorem proving strategies that exploit the structure of a KB to improve the efficiency of reasoning. We analyze and experimentally evaluate these strategies with a testbed based on the SNARK theorem prover. Exploiting graph-based partitioning algorithms, we show how to compute a partition-derived ordering for ordered resolution, with good experimental results, offering an automatic alternative to hand-crafted orderings. We further propose a new resolution strategy, MFS resolution, that combines partition-based message passing with focused sublanguage resolution. Our experiments show a significant reduction in the number of resolution steps when this strategy is used. Finally, we augment partition-based message passing, partition-derived ordering, and MFS by combining them with the set-of-support restriction. While these combinations are incomplete, they often produce dramatic improvements in practice. This work presents promising practical techniques for query answering with large and potentially distributed commonsense KBs.","doi":null,"url":"https://ijcai.org/proceedings/2003/13","pdf":"https://ijcai.org/proceedings/2003/0013.pdf","markdown":"https://ijcai.org/proceedings/2003/13.md","bibtex":"https://ijcai.org/proceedings/2003/bibtex/13","volume":{"slug":"2003","title":"Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2003","url":"https://ijcai.org/proceedings/2003/","year":2003,"location":"Acapulco, Mexico","dates":"","editors":"Georg Gottlob","isbn":"","publisher":"International Joint Conferences on Artificial Intelligence","doi_prefix":""},"section":null,"track":"","number":13,"pages":{"first":89,"last":97},"keywords":[],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}