World-wide Intelligent Semantic
Patent Extraction and Retrieval
Funding programme: 5th Framework Programme – IST
Star – end year: 2002 – 2005
Partnership
British Maritime Technology, United Kingdom
HAPTICA Limited, Ireland
SYNTHEMA, Italy
TEMIS, France
TXT e-solutions
Consorzio Milano Ricerche
The project
The next generation Internet requires that digital content be augmented to fulfil the needs of intelligent retrieval systems. Content needs some sense of semantics such that it is capable of processing by multiple agents and can be re-used in multiple contexts. The project have been tackling a familiar type of structured content, namely patents, with the aim of demonstrating the development of flexible tools that permit their automated mark-up.
The subject matter of patents themselves is widespread, but the way in which patents are used and how they are referred to is not. We are in position to capture these regularities in both content and man-machine interaction so that patent content becomes much less opaque and suitable for directed search and retrieval.
Thus, WISPER is a new multilingual portal for intelligent access to massive patent databases. It encodes and visualises semantic content of patents by using advanced text mining, intelligent image analysis and user profiling to allow ontology development and thereby encourage SME participation in patenting. Users will be able to search both text and images for the first time, and results will be shown as detailed graphical interfaces.