Altshuller’s matrix, over various conducted surveys on the frequency of use by practitioners, remains systematically in the lead despite criticism of its obsolescence. Consequently, attempts have emerged to update it, both in terms of principle’s quantity and their statistical distribution in generic technical conflicts. Nevertheless, up to now, none of them has supplanted the effectiveness of the Altshuller matrix. These attempts as well as other approaches introducing new tools for patent classification and information retrieval often suffer from poor accuracy in data extraction. In this paper, we introduce a new TRIZ-dedicated extraction tool based on a deep neural network summarization called SummaTRIZ. We also introduce a method including SummaTRIZ to update TRIZ matrix and create a whole new matrix based on patents and independent from potentially obsolete inventive principles.
CITATION STYLE
Guarino, G., Samet, A., & Cavallucci, D. (2020). Summarization as a Denoising Extraction Tool. In IFIP Advances in Information and Communication Technology (Vol. 597 IFIP, pp. 77–87). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-61295-5_7
Mendeley helps you to discover research relevant for your work.