Limiting logical violations in ontology alignment through negotiation



Jimenez Ruiz, E, Payne, TR ORCID: 0000-0002-0106-8731, Solimando, A and Tamma, VAM ORCID: 0000-0002-1320-610X
(2016) Limiting logical violations in ontology alignment through negotiation. In: 15th International Conference on Principles of Knowledge Representation and Reasoning, 2016-04-25 - 2016-04-29, Cape Town, South Africa.

[img] Text
kr-2016-crc.pdf - Unspecified

Download (572kB)

Abstract

Ontology alignment (also called ontology matching) is the process of identifying correspondences between entities in different, possibly heterogeneous, ontologies. Traditional ontology alignment techniques rely on the full disclosure of the ontological models; however, within open and opportunistic environments, such approaches may not always be pragmatic or even acceptable (due to privacy concerns). Several studies have focussed on collaborative, decentralised approaches to ontology alignment, where agents negotiate the acceptability of single correspondences acquired from past encounters, or try to ascertain novel correspondences on the fly. However, such approaches can lead to logical violations that may undermine their utility. In this paper, we extend a dialogical approach to correspondence negotiation, whereby agents not only exchange details of possible correspondences, but also identify potential violations to the consistency and conservativity principles. We present a formal model of the dialogue, and show how agents can repair logical violations during the dialogue by invoking a correspondence repair, thus negotiating and exchanging repair plans. We illustrate this opportunistic alignment mechanism with an example and we empirically show that allowing agents to strategically reject or weaken correspondences when these cause violations does not degrade the effectiveness of the alignment computed, whilst reducing the number of residual violations.

Item Type: Conference or Workshop Item (Unspecified)
Additional Information: Copyright © 2016, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. This article has been accepted for the 15th International Conference on Principles of Knowledge Representation and Reasoning (2016). About The Authors
Depositing User: Symplectic Admin
Date Deposited: 06 May 2016 08:17
Last Modified: 25 Aug 2022 07:10
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3001089