alekstoteles/NLP_final_project

Natural Language Processing with Deep Learning (Fall 2019)

Sparse Multi-label Patent Classification with Deep Learning

Alexander Mueller & Kevin Stone

W266 NLP with Deep Learning - Final Project (Fall 2019)

Project description

The project developed a multi-label classification model capable of categorizing patents into Cooperative Patent Classification (CPC) codes. Classifications are done on the CPC sub-class level and are based on patent abstract texts from the USPTO 2M dataset.

Files

• Final Paper (pdf)

• Final Presentation (pdf)

• Code

External links

• Link to USPTO 2M dataset

Acknowledgements

We want to thank Joachim Rahmfeld (UC Berkeley, School of Information) for his guidance on our project.