English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This study provides an efficient approach for using text data to calculatepatent-to-patent (p2p) technological similarity, and presents a hybridframework for leveraging the resulting p2p similarity for applications such assemantic search and automated patent classification. We create embeddings usingSentence-BERT (SBERT) based on patent claims. We leverage SBERTs efficiency increating embedding distance measures to map p2p similarity in large sets ofpatent data. We deploy our framework for classification with a simple NearestNeighbors (KNN) model that predicts Cooperative Patent Classification (CPC) ofa patent based on the class assignment of the K patents with the highest p2psimilarity. We thereby validate that the p2p similarity captures theirtechnological features in terms of CPC overlap, and at the same demonstrate theusefulness of this approach for automatic patent classification based on textdata. Furthermore, the presented classification framework is simple and theresults easy to interpret and evaluate by end-users. In the out-of-sample modelvalidation, we are able to perform a multi-label prediction of all assigned CPCclasses on the subclass (663) level on 1,492,294 patents with an accuracy of54% and F1 score > 66%, which suggests that our model outperforms the currentstate-of-the-art in text-based multi-label and multi-class patentclassification. We furthermore discuss the applicability of the presentedframework for semantic IP search, patent landscaping, and technologyintelligence. We finally point towards a future research agenda for leveragingmulti-source patent embeddings, their appropriateness across applications, aswell as to improve and validate patent embeddings by creating domain-expertcurated Semantic Textual Similarity (STS) benchmark datasets.

PatentSBERTa: A Deep NLP based Hybrid Model for Patent Distance and Classification using Augmented SBERT