Released "Camphr", a natural language processing library, as open source.
- Natural Language Processing
Overview
- Project period
- February 12, 2020 (OSS release)
We have developed and released "Camphr" with the aim of increasing the efficiency and sophistication of research and service development of natural language processing systems by researchers and engineers, and thereby accelerating the implementation of natural language processing technology in society.
Camphr is implemented as a plug-in for spaCy, a framework that enables unified handling and pipelining of various methods.
Using Camphr, developers can easily combine state-of-the-art deep learning methods and conventional algorithms using patterns, dictionaries, etc., within the framework of spaCy.
Technology
Camphr is open source software for natural language processing, allowing any combination of state-of-the-art deep learning methods and conventional algorithms using patterns, dictionaries, etc.
It is implemented as a plug-in for spaCy, a natural language processing framework, and facilitates the assembly and handling of a series of processes as a pipeline.
Camphr provides language analysis and learning functions using deep learning models such as BERT and XLNet, multilingual support, keyword extraction, embedded vector arithmetic, and parsing functions using KNP.
Member in charge
WATANABE yotaro
D. from Nara Institute of Science and Technology (NAIST), and Assistant Professor at Graduate School of Information Science and Technology, Tohoku University.
During that time, he was mainly engaged in research on natural language processing, especially semantic analysis and inference techniques of language.
After working as a project leader of AI-related technologies at a major electronics manufacturer, he joined PKSHA Technology. D. in Engineering.
He is a former board member of the Japan Deep Learning Association.
