{"id":467,"date":"2021-12-10T11:16:22","date_gmt":"2021-12-10T10:16:22","guid":{"rendered":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/nouveausite\/?page_id=467"},"modified":"2022-01-11T11:32:51","modified_gmt":"2022-01-11T10:32:51","slug":"these-rinaldo-lima","status":"publish","type":"page","link":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/these-rinaldo-lima\/","title":{"rendered":"These Rinaldo Lima"},"content":{"rendered":"

\"\"Rinaldo Lima<\/span><\/strong>, \u00ab\u00a0OntoILPER: An Ontology and Inductive Logic Programming-based Method to Extract Instances of Entities and relations from Texts\u00a0\u00bb (defended in 2014).<\/span><\/p>\n

Manuscrit<\/a><\/span><\/p>\n

La th\u00e8se de Rinaldo Lima (soutenue en 2014), en collaboration avec l’universit\u00e9 F\u00e9d\u00e9rale du Pernambuco (UFPE), porte sur l’extraction supervis\u00e9e d’entit\u00e9s nomm\u00e9es (NER) et de relations entre ces entit\u00e9s. Nous avons propos\u00e9 une m\u00e9thode supervis\u00e9e pour extraire des instances d’entit\u00e9s et de relations de corpus textuels bas\u00e9 sur la programmation logique inductive (une technique d’apprentissage machine symbolique) et exploitant des ressources s\u00e9mantiques, notamment des ontologies. Cette m\u00e9thode propos\u00e9e s’appuie aussi sur un espace d’hypoth\u00e8se relationnelle tr\u00e8s expressif, sous la forme de pr\u00e9dicats logiques conduisant \u00e0 une repr\u00e9sentation du contenu des documents sous la forme d’un graphe permettant \u00e9galement l’exploitation d’une ontologie de domaine et de connaissances complementaires notamment lexicale, syntaxique, s\u00e9mantique et relationnelles. A partir de cette repr\u00e9sentation un prototype nomm\u00e9 OntoILPER a \u00e9t\u00e9 d\u00e9velopp\u00e9 mettant en oeuvre la programmation logique inductive (PLI) et induisant automatiquement des r\u00e8gles symboliques d’extraction d’information (entit\u00e9s et relations binaires) en Prolog traduites ensuite en SWRL.\u00a0 Ces r\u00e8gles sont ensuite appliqu\u00e9es sur des documents qui n’ont pas servi \u00e0 l’apprentissage et permettent de peupler (populer) l’ontologie de domaine. Les r\u00e9sultats obtenus par OntoILPER sur des corpus de r\u00e9f\u00e9rences ont d\u00e9montr\u00e9 son efficacit\u00e9, tant en extration d’entit\u00e9s nomm\u00e9es que de relations binaires entre ces entit\u00e9s, par rapport \u00e0 d’autres syst\u00e8mes mettant en oeuvre des techniques d’extraction statistiques.<\/span><\/p>\n

The Rinaldo Lima thesis (defended in 2014), in collaboration with the Federal University of Pernambuco (UFPE), concerns supervised named entity recognition (NER) and relationships extraction (RE) between these entities . We have proposed a method to extract supervised entity instances and text corpora relationships based on inductive logic programming (a learning technique symbolic machine) and exploiting semantic resources such as ontologies. The proposed method is also based on a very expressive space relational hypothesis in the form of predicate logic leading to a representation of the content of the documents in the form of a graph also allows the operation of a domain ontology of particular lexical complementary knowledge, syntactic, semantic and relational. From this representation a prototype named OntoILPER was developed utilizing inductive logic programming (ILP) and automatically inducing symbolic information extraction rules (binary entities and relations) in Prolog then translated into SWRL. These rules are then applied on documents that were not used to learning and enable people (populate) the domain ontology. The results obtained by OntoILPER on corpora of references have demonstrated efficacy in both extration named entity as binary relations between these entities, compared to other systems by implementing statistical extraction techniques.<\/em><\/span><\/p>\n

\"\"<\/p>\n","protected":false},"excerpt":{"rendered":"

Rinaldo Lima, \u00ab\u00a0OntoILPER: An Ontology and Inductive Logic Programming-based Method to Extract Instances of Entities and relations from Texts\u00a0\u00bb (defended in 2014). Manuscrit La th\u00e8se de Rinaldo Lima (soutenue en 2014), en collaboration avec l’universit\u00e9 F\u00e9d\u00e9rale du Pernambuco (UFPE), porte sur l’extraction supervis\u00e9e d’entit\u00e9s nomm\u00e9es (NER) et de relations entre ces entit\u00e9s. Nous avons propos\u00e9 … <\/p>\n

Continuer la lecture \u00ab\u00a0These Rinaldo Lima\u00a0\u00bb<\/span><\/a><\/p>\n","protected":false},"author":2,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_crdt_document":"","footnotes":""},"class_list":["post-467","page","type-page","status-publish","hentry","entry"],"_links":{"self":[{"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/pages\/467","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/comments?post=467"}],"version-history":[{"count":13,"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/pages\/467\/revisions"}],"predecessor-version":[{"id":1227,"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/pages\/467\/revisions\/1227"}],"wp:attachment":[{"href":"https:\/\/pageperso.lis-lab.fr\/bernard.espinasse\/index.php\/wp-json\/wp\/v2\/media?parent=467"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}