Early research on machine learning in artificial systems date back to the 1950's. Fromthe 80 start to develop practical applications of algorithms called "subsymbolic"(mainly Bayesian neural networks and systems) to problems of pattern recognitionand classification and the "symbolic" (induction of trees and rules) to knowledge acquisition for expert systems. In the 90 off with force what has been called 'data mining' application of learning algorithms and visualization for knowledge extractionin large databases. Our contributions in this field have also followed this path:
This field works in the development of tools for word processing in Spanish (ARIESprojects, CRATER, ATILA and accordion). These projects have developed:
Lexical resources: ARIES platform, with more than 40,000 slogans in Spanish,covering more than 500,000 inflected forms.
Tools: Libraries of tools for building applications. In particular, access and efficient management of large volumes lexical, morphological analyzers, taggers, parsers,etc.
Applications: constructed from the available tools. These include applications forindexing and retrieval, spell checking, style checking, etc.
These topics are collaborating with the Laboratory of Computational Linguistics at the Autonomous University of Madrid.
In this line of work, the group is researching and implementing methodologies andtechnologies that enable agile development of applications, including Web services, enhancing reliability and safety.
The "semantic web" is being developed with the aim to define and link Web resources so that programs (agents) can interpret not only for presentation purposes but for automation, integration and reuse. This makes it possible to offer on the new network infrastructure and more efficient services. This requires accurately describe the resources and the relationships between them, ie formally represent the knowledge inherent in the domain that are a part. Work on knowledge representation, classic artificial intelligence, have derived the "ontology engineering", which deals with methodologies for the development of 'ontology' formal representations of concepts and relationships in a domain.Results 1 - 6 of 6