A Multi-task Approach to Neural Multi-label Hierarchical Patent Classification Using Transformers

Pujari, Subhash Chandra; Friedrich, Annemarie; Strötgen, Jannik

doi:10.1007/978-3-030-72113-8_34

Cited by 4 publications

(4 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Comparing to the neural baseline, TMM with e(t + a), which has recently reported state-ofthe-art results on CPC classification (Pujari et al, 2021), we find that adding information from additional text fields and the CPC embeddings consistently improves performance.…”

Section: Comparison With the Baselinesmentioning

confidence: 66%

“…Patent Datasets. For CPC classification, various datasets are available (Pujari et al, 2021(Pujari et al, , 2022Li et al, 2018). Similar to our target classification task, Richter and MacFarlane (2005) study classification for a patent alert system for the biochemical domain, but the dataset was not open-sourced.…”

Section: Related Workmentioning

confidence: 99%

“…CNNs (Li et al, 2018;Niu and Cai, 2019), RNNs (GRU Krestel, 2019), andLSTMs (Grawe et al, 2017)) have also been used to represent patent text. Recently, pre-trained transformer-based models have been shown to be effective for patent classification (Lee and Hsiang, 2020;Pujari et al, 2021;Althammer et al, 2021). Transformer-based models are constrained to a maximum length of 512 input tokens.…”

Section: Related Workmentioning

confidence: 99%

“…Most prior work on patent classification engages in the task of assigning labels from the International Patent Classification (IPC) taxonomy and its extension, the Co-operative Patent Classification (CPC) taxonomy (Smith, 2002;Grawe et al, 2017;Li et al, 2018;Lee and Hsiang, 2020;Pujari et al, 2021). This hierarchical multi-label classification task is challenging due to the large number of CPC codes, but lots of training data exists as each patent gets assigned CPC labels upon submission to the patent office.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Three Real-World Datasets and Neural Computational Models for Classification Tasks in Patent Landscaping

Pujari¹,

Strötgen²,

Giereth³

et al. 2022

Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

Patent Landscaping, one of the central tasks of intellectual property management, includes selecting and grouping patents according to userdefined technical or application-oriented criteria. While recent transformer-based models have been shown to be effective for classifying patents into taxonomies such as CPC or IPC, there is yet little research on how to support real-world Patent Landscape Studies (PLSs) using natural language processing methods. With this paper, we release three labeled datasets for PLS-oriented classification tasks covering two diverse domains. We provide a qualitative analysis and report detailed corpus statistics.Most research on neural models for patents has been restricted to leveraging titles and abstracts. We compare strong neural and non-neural baselines, proposing a novel model that takes into account textual information from the patents' full texts as well as embeddings created based on the patents' CPC labels. We find that for PLS-oriented classification tasks, going beyond title and abstract is crucial, CPC labels are an effective source of information, and combining all features yields the best results.

show abstract

Section: Comparison With the Baselinesmentioning

confidence: 66%