Using patents to populate an inventive design ontology

AbstractTRIZ theory is largely based on patent's observation, and as we’ve learned from TRIZ, key issues (amongst others) like automatic contradiction extraction, accurately estimate a patent's level of Inventiveness, are still to be efficiently addressed. An evident bottleneck lies in the necessary cross disciplinary research to be undertaken between data/text mining algorithms, Design theories and methodologies and psychological aspects of creativity. This paper is firstly proposing to summarize key contributions and major research bottlenecks in patents usage in the scope of TRIZ-related practices. Then a positioning between TRIZ and our Inventive Design Method is exposed. A third part reports on our contribution, it consist in a framework aiming at extracting and representing know-how of domain experts and populating an already constructed Ontology of Inventive Design. This framework is using the notion of problem graph and polycontradictions. All these processes are discussed in order to interface with our software prototype: TRIZAcquisition