Meaningful patent searches
The envisioned solution, the goal of the 'Tool platform for intelligent patent analysis and summarization' (TOPAS) project, will support patent specialists’ daily work, enhance their performance and reduce costs for patent tasks. Work on the development of TOPAS technologies builds on results achieved in the Sixth Framework Programme project 'Advanced patent document processing techniques' (Patexpert). To achieve project goals, consortium members have outlined a set of topics key to their approach: named entity recognition, segmentation, term detection, relation extraction and summarisation. To date, TOPAS has carried out the study and annotation of patent documentation, and compiled an initial balanced patent corpus. A database has also been created to store the data. Goals related to work on the development of technologies for patent material analysis have been fully achieved, with the basic named entity recognition system having been implemented. This supports basic terminology finding and facilitates classification for patent citations and physical measurements. TOPAS has also completed the design of a refined hierarchical scheme for segmentation and labelling of patents, which is the basis for advanced segmentation, labelling and summarisation. Partners have also completed all objectives relevant to the development of technologies for patent material summarisation. Other project achievements include the compilation of the patent corpus and setup of the workbench's computational infrastructure. Ongoing project activities relate to validation by pilot use case realisation. This will commence once the already developed technologies are in place and partially integrated into the workbench. By project end, TOPAS expects to have delivered a comprehensive software platform for the semantic analysis and summarisation of patents. Project outcomes are slated to make a significant contribution to the competitiveness of innovation-oriented companies with an intellectual property rights agenda. This translates to benefits for the clients of small and medium-sized enterprises' products, i.e. patent departments within companies, patent lawyers and innovators. All TOPAS technologies will support multilingual patent documentation in the languages of the European Patent Office: German, French and English.
Keywords
Patents, patent search, content analysis, patent corpus, material summarisation
