Read out

Guest Talk: Semantic Interactive Ontology Matching: Synergistic Combination of Techniques to Improve the Set of Candidate Correspondences

Dr. Kate Revoredo 

Date/Time: 25.10.2017, 17:00 

Loca­tion: D2.2.094 

Abstract 

Onto­logy Matching is the task of finding a set of entity corre­spon­dences between a pair of onto­lo­gies, i.e. an align­ment. It has been recei­ving a lot of atten­tion due to its broad appli­ca­tions. Many tech­ni­ques have been proposed, among which the ones applying inter­ac­tive stra­te­gies. An inter­ac­tive onto­logy matching stra­tegy uses expert know­ledge towards impro­ving the quality of the final align­ment. When these stra­te­gies are based on the expert feed­back to vali­date corre­spon­dences, it is important to esta­blish criteria for selec­ting the set of corre­spon­dences to be shown to the expert. A bad defi­ni­tion of this set can prevent the algo­rithm from finding the right align­ment or it can delay conver­gence. On the other way around, tech­ni­ques that consider the seman­tics of the enti­ties of the onto­logy have shown good results. In this work we present tech­ni­ques which, when used simul­ta­neously, improve the set of candi­date corre­spon­dences. These tech­ni­ques are incor­po­rated in an inter­ac­tive onto­logy matching approach, called ALINSem. Expe­ri­ments success­fully show the poten­tial of our proposal.

Bio 

Kate Revoredo is an Asso­ciated Professor of the Depart­ment of Applied Infor­ma­tics at the Federal Univer­sity of the State of Rio de Janeiro (UNIRIO), Brazil. She obtained a D.Sc. and a M.Sc. in Computer Science with emphasis in Arti­fi­cial Intel­li­gence from the Federal Univer­sity of Rio de Janeiro (COPPE-UFRJ). During her D.Sc. studies in the context of auto­matic adap­ta­tion of proba­bi­listic rela­tional models, she was a visiting rese­ar­cher at Machine Learning and Natural Language Proces­sing Lab at Alber­t-Lud­wigs­-­Uni­ver­sity Frei­burg, Germany. Her rese­arch focus is mainly on machine learning and data mining, more speci­fi­cally on algo­rithms for learning and adap­ting onto­lo­gies and their align­ments through data. More­over, she is also work with process disco­very and moni­to­ring through data. She has published in important jour­nals and confe­rence papers, parti­ci­pates in several program commit­tees of jour­nals and confe­rences, and is a member of the Brazi­lian Computer Society and the Brazi­lian Special Commis­sion in Arti­fi­cial Intel­li­gence.



Back to overview