CNN based Estimating Patent Quality: Focusing on AI Convergences

Recently, obtaining high quality patents has become increasingly important for the pursuit of technological convergence. However, it is insufficient for effectively estimating patent quality, especially in the emerging convergent technology. This research newly proposed a technique to evaluate patents, focusing on AI convergences, by considering both bibliometric aspects and embedded citations of patents with Convolutional Neural Network. Findings of this research empirically contribute to the estimation of patent quality. The proposed technique is expected to discover patents with high quality. Related policy implications based on this research could leverage R&D management in AI area.