WordNet: a lexical database for English
Communications of the ACM
Learning Information Extraction Rules for Semi-Structured and Free Text
Machine Learning - Special issue on natural language learning
SIGDOC '86 Proceedings of the 5th annual international conference on Systems documentation
Active Learning for Natural Language Parsing and Information Extraction
ICML '99 Proceedings of the Sixteenth International Conference on Machine Learning
Bootstrapping an ontology-based information extraction system
Intelligent exploration of the web
Automatic association of web directories with word senses
Computational Linguistics - Special issue on web as corpus
Unsupervised word sense disambiguation rivaling supervised methods
ACL '95 Proceedings of the 33rd annual meeting on Association for Computational Linguistics
Information retrieval using word senses: root sense tagging approach
Proceedings of the 27th annual international ACM SIGIR conference on Research and development in information retrieval
Evaluating WordNet-based Measures of Lexical Semantic Relatedness
Computational Linguistics
CRYSTAL inducing a conceptual dictionary
IJCAI'95 Proceedings of the 14th international joint conference on Artificial intelligence - Volume 2
Hi-index | 0.00 |
In this paper, by considering a particular application field, the innovation, we propose an automatic system to feed an innovation knowledge base (IKB) starting from texts located on the Web. To facilitate the extraction of concepts from texts we distinguished in our work two knowledge types: primitive knowledge and definite knowledge. Each one is separately represented. Primitive knowledge is directly extracted from natural language texts and temporally organized in a specific base called TKB (Temporary Knowledge Base). The entry of the base IKB is the knowledge filtered from the TKB by some specified rules. After each filtering step, the TKB is emptied for starting new extractions from other texts sources. The filtering rules are specified using variables representing interesting concepts. Their specifications result from the semantics of the innovation operators involved in the innovation process. The variables are initiated from a semantic representation of the operators. The content of the base IKB can be displayed as text annotations. Hence the feeding system is coupled with a user interface allowing the exploration of these annotations through their dynamic insertion in the associated texts. In this paper, we present the application field and our approach for representing and for feeding the IKB innovation base. We also provide a number of experiment results and we indicate work we plan to undertake in order to improve our system.