Find the most recent information on EU Funding activities in the field of Information and Communication Technologies (ICT) by visiting our ICT in FP7 website, which covers ICT in the 7th Framework Programme (FP7) 2007 - 2013.
PATExpert
Advanced Patent Document Processing Techniques
PATExpert developed a multimedia content representation formalism based on Semantic Web principles for selected technology areas and investigated new generation semantic patent processing techniques.
Impact
PATExpert is about to shift the paradigm for patent documentation processing from textual (viewing patents as text blocks enriched by “canned” picture material, sequences of morpho-syntactic tokens, or collections of syntactic structures) to semantic (viewing patents as multimedia knowledge objects). In the long term, this will essentially change the way patents are authored, filed, stored and processed.
Main innovation
The focus of PATExpert on content (rather than on surface) is expected to lead to a number of innovative technologies in the following areas:
- Semantic web representation mechanisms for patent material: PATExpert will provide a content and meta-content representation formalism and put it to use by building up knowledge bases (KBs) for two sample patent technology areas: optical recording and machine tools.
- User-oriented patent classification: For a user, it is essential to be able to define upon the International Patent Classification a dynamic classification schema that is based on his own criteria. PATExpert will facilitate the creation and maintenance of dynamic classification schemas and provide the means for automatic classification of patents in accordance with these schemas.
- Patent information / and meta-information extraction and (semi-) automatic construction of patent knowledge bases: PATExpert will develop shallow and deep techniques for the distillery of content from patent documentation and its representation in a KB.
- Semantic multimodal patent retrieval: Along with keyword-based retrieval, PATExpert will work on new generation retrieval techniques that are based on patent text and image similarity, on content, and on meta-information.
- Patent text paraphrasing: Patent material is very difficult to read due to its very complex linguistic constructions. Shallow and deep techniques for paraphrasing of patent documentation developed in PATExpert will lead to an improved and considerably more readable style.
- Multilingual patent gist (summary) generation: Users are usually interested in certain specific information contained in a patent (application) rather than in a patent (application) as a whole. The techniques developed in PATExpert will facilitate the generation of concise summaries that contain just the desired information in the preferred language of the user.
- Intelligent navigation in patent spaces and their visualization: The objects in complex patent representation spaces (be they DBs or KBs) are related by a multitude of different content and meta information links. PATExpert will develop intelligent techniques for visualizing these links and support navigation along them.
- Patent valuing and technology area assessment: Patent valuing and assessment of technology areas with respect to their potential to give rise to patent applications is a highly demanding issue. PATExpert will contribute to the field by developing techniques that make use of semantic criteria.
Results
PATExpert started in February 2006. The first two major tasks consisted in the assessment of the state of the art in all areas of patent processing addressed within PATExpert and the development of the first version of the semantic representation formalism for patent documentation. This formalism forms the foundation of the work on all other topics of PATExpert. In parallel, the first version of the domain ontology for the sample area of optical recording has been developed.
More details
- Annual Report 2008 (PDF, 429KB)
- Annual Report 2007 (PDF, 459KB)
- Annual Report 2006 (PDF, 317KB)
- Project Presentation (PDF, 179KB)
- Flyer (PDF, 3.386KB)
- ICT Results article (including Video): Faster, better patent processing (16 April 2009) Processing patents is complex and time-consuming. Using semantic web technology, researchers have come up with a powerful tool to process patents faster and better.
- PATExpert project web site
Administrative details
- PATExpert (IST FP6 028116) is a Specific Targeted Research Project under the 4th call of the European Union's 6th Framework Programme, Thematic Priority 2 (Information Society Technologies), where it addresses the Strategic Objective “Semantic-Based Knowledge Systems”.
- The project started on 1 February 2006 and finished on 31 July 2008; it involved eight partners from four different European countries.
- PATExpert's overall budget is about 3.800.000 €
List of participants
- FBM, Pompeu Fabra University, Spain (coordinator)
- University of Stuttgart, Germany
- Aristotle University of Thessaloniki, Greece
- Il Centro per la Ricerca Scientifica e Tecnologica, Italy
- Industrie Software Jochen Brägmann, Germany
- Fraunhofer Gesellschaft, Germany
- European Patent Office, Germany
- IALE, Spain
Contact persons
- Coordinator
- Leo Wanner
- FBM, Pompeu Fabra University
- Administrative and financial management
- Eugenia Fuenmayor
- FBM, Pompeu Fabra University
Events in connection with PATExpert
- Presentation of the paper 'Application of Semantic Technologies for Representing Patent Metadata' (PDF, 196KB), Informatik October 2006, Dresden, Germany
- Presentation of the paper 'PATExpert: Semantic Processing of Patent Documentation' (PDF, 57KB) Project Session, SAMT 2006, December 2006, Athens, Greece