A Label Informative Wide & Deep Classifier for Patents and Papers

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 3438–3443, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 3438 A Label Informative Wide & Deep Classifier for Patents and Papers Muyao Niu Patsnap Ltd. niumuyao@patsnap.com Jie Cai Patsnap Ltd. caijie@patsnap.com Abstract In this paper, we provide a simple and effective baseline for classifying both patents and pa- pers to the well-established Cooperative Patent Classification (CPC). We propose a label- informative classifier based on the Wide & Deep structure, where the Wide part encodes string-level similarities between texts and la- bels, and the Deep part captures semantic- level similarities via non-linear transforma- tions. Our model trains on millions of patents, and transfers to papers by developing distant- supervised training set and domain-specific features. Extensive experiments show that our model achieves comparable performance to the state-of-the-art model used in industry on both patents and papers. The output of this work should facilitate the searching, granting and filing of innovative ideas for patent exam- iners, attorneys and researchers. 1 Introduction Classifying patents and papers to a technology taxonomy is a crucial step to organize the mas- sive knowledge and to discover innovative ideas. Patent examiners rely on the taxonomy to search for similar documents when granting or invali- dating a patent application; attorneys use it to check whether the innovation points of an inven- tion have been covered in previous literature; re- searchers use the taxonomy to monitor the tech- nology trends in certain fields, and companies use it to outline the intellectual property landscape of its own or its competitors’. The most commonly used taxonomies are International Patent Classifi- cation (IPC) 1 and its newer version Cooperative Patent Classification (CPC) 2. Figure 1 illustrates the CPC hierarchy and the discriminative descrip- tions attached to each node. 1https://www.wipo.int/classifications/ipc/en/ 2https://www.epo.org/searching-for-patents/helpful- resources/first-time-here/classification/cpc.html Figure 1: Illustration of the CPC hierarchy. The categorization is now mostly done manu- ally by experts in patent offices. Due to the grow- ing amount of patents and the limited number of domain experts, there has been an urge to auto- mate the classification process. Also, as more and more technology innovations are published in patents, researchers with no background in patent classification may want to know which patents are most relevant to an academic paper. For this end, we aim to classify both patents and papers to the CPC subclass with more than 600 labels. Classifying patents and papers to CPC subclass is a challenging task because (1) there are a large number of labels that covers almost all technology domains, and the differences between labels are often subtle; (2) although mass amount of labelled data for patents is available, the annotated data for paper-to-CPC is very limited. Since labelling pa- pers with CPC labels requires expert knowledge, large-scale human annotation is very expensive. In this paper, we leverage the CPC label de- scriptions and use the Wide-and-Deep network to integrate label information with …

Content truncated to 500 words