An AI research institute specializing in natural language processing research, established by Works Applications in Tokushima Prefecture in February 2017.

HUE, the cloud-based ERP package software developed by Works, utilizes operation log data accumulated within a company as training data for machine learning.

The institute is conducting research specifically on business applications of natural language processing in order to commercialize AI functions that better meet user needs based on the analysis and understanding of user speech and text data.

What is Natural Language Processing (NLP)?

Natural language processing (NLP) is a general term for technologies and software that allow computers to process human-used languages such as Japanese and English, and is a branch of artificial intelligence and linguistics.
NLP is used in all kinds of situations, from kana-kanji conversion for keyboard input to fuzzy search functions and voice recognition functions.

Activities

The Institute has released open source NLP tools and language resources (morphological analysis "Sudachi") to Japan and the English-speaking world. We also present our research at conferences, workshops, and seminars.

Release of the morphological analyzer OSS "Sudachi

In August 2017, we released the morphological analyzer "Sudachi" as open software.

Sudachi" is a morphological analyzer that combines flexibility and accuracy lacking in existing morphological analyzers with high-quality analysis performance that can withstand commercial use.
Sudachi" can be used for a variety of analysis processes, such as reading text strings in images and PDFs, and for chatbots that can obtain necessary information in a conversational style using an AI-based automatic conversation program.

By improving the accuracy of "Sudachi" and expanding the scope of its use, we aim to realize AI functions that fit the operations of a wide variety of companies.

In May 2008, the Ministry of Economy, Trade and Industry (METI) established the "Study Group for Digital Transformation (DX)" to organize the current issues and consider measures to address them in order for Japanese companies to realize DX, focusing on the state of IT systems, and compiled a DX report.

The report recommended the need to establish guidelines and DX promotion indicators so that there is shared recognition of the approaches and necessary actions to realize DX, and the initiative has begun to take off.

[Resource] DX Report - Overcoming the IT System "2025 Cliff" and Full-Scale Deployment of DX

IPA is involved in the revision of the above guidelines for DX, the creation of evaluation indicators related to IT indicators, and the development of benchmarks, and our employees have been seconded to IPA to participate in this project.

In the future, the public and private sectors will continue to work together, and each company will need to restructure its IT systems, acquire and train DX personnel, etc. for DX.
We have proposed business improvement and efficiency enhancement to more than 2,000 clients without customization for each individual company.

We would like to contribute to the promotion of DX for Japanese companies with our Products/Services that incorporate and update many customers' requests without customization, which causes black boxes, and generalize "standard functions" based on our know-how and the various business requirements and business practices required by companies regardless of industry or business category. We are committed to contributing to the promotion of DX among Japanese companies.