Abstract
Industrial technology matching events are held by governmental institutions worldwide to promote patent transfer from universities to industries. When collecting academic patents for the matching events, governmental institutions lack professional knowledge for identifying academic patents suitable for various industries. Therefore, previous studies adopted International Patent Classification (IPC) codes assigned by patent examiners to represent patents and mined the industry-related cues through the mapping link between IPC codes and industry categories. However, IPC codes are too general to specifically represent the complex patents, leading to inaccurate tagging. The view of patent inventors (e.g., patent titles and abstracts) contains rich industry-related cues that benefit assigning industrial categories to academic patents. Therefore, we propose a dual-view attention neural network that learns low-dimensional patent representations from the views of patent examiners and inventors and merges the representations for classifying academic patents into suitable industrial categories. Experiments show that the proposed method outperforms benchmark methods.
Recommended Citation
Liu, Zhaobin; Deng, Weiwei; Zhu, Peihu; DU, Wei; and Ma, Jian, "A Dual-view Attention Neural Network for Assigning Industrial Categories to Academic Patents" (2023). PACIS 2023 Proceedings. 73.
https://aisel.aisnet.org/pacis2023/73
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
Paper Number 1364; Track Design; Complete Paper