On July 2020, 7, FRONTEO co-sponsored Retrieva Inc. and Nextremer Inc., and held an online seminar with the seminar title "The management of a natural language processing vendor talks! The key points of text data analysis that can be used from tomorrow." ..

FRONTEO x Retriever x Nextremer co-sponsored seminar Cross talk scene (from left to right, Retriever President Kazuya Kawahara, FRONTEO CTO Hideki Takeda, Nextremer Chairman Nagahiro Mukai, FRONTEO AI BizDevOps Team Manager Kenya Igasaki II)

From FRONTEO, Director and CTO Hideki Takeda will be on stage, with the title of the lecture "Learning from the introduction examples of more than 200 companies, equations that generate business value from natural language", and at the beginning "Collaboration between AI and humans" and "(in the popularization period) I told you about "new common sense of AI".In particular, he mentioned "application" as an important point that was introduced in the new common sense, and emphasized the need to make it usable in the work of a specific job.Then, he explained the equation "data x learning-> people x work" with ideas about problem-solving processes based on data, and finally talked about KIBIT utilization examples that practice them.

Our CTO Takeda giving a lecture

Mr. Kazuya Kawahara, President and CEO of Retriever Co., Ltd., gave a lecture entitled "What is AI that analyzes text data into'Yoshina'" and promoted understanding of the difficulty of analyzing text data and the approach using AI. ..We explained a concrete example of using our own analysis AI "YOSHINA" for text data whose analysis axis is difficult to understand compared to numerical data whose analysis axis is easy to understand.

Mr. Nagahiro Mukai, Chairman and CEO of Nextremer Co., Ltd., said, "The key technology of the WITH corona era," remote ", is the theme of the decrease in the working population, changes in communication methods due to the influence of corona, and sophistication and efficiency of services by DX. He said that "remote" will be a key technology in the future, and introduced examples of face-to-face operations such as condominium management operations and counter operations, using the example of a multi-modal dialogue system that utilizes the company's AI.

In the subsequent crosstalk sessions at the three companies, Kenya Igasaki, General Manager of the FRONTEO AI BizDevOps Team, served as the moderator and held a panel discussion on the current status and future of practical application of the natural language processing industry. In the content that talked about the strengths of each company while looking at the item of natural language processing on Wikipedia, the interest of OCR and the utilization of time series information were picked up among the items of natural language processing.

In addition, FRONTEO's Takeda mentioned "deep dive domain" in the theme of "initiatives and research areas that each company has been paying attention to recently".He said that it is important to make an application that can be used in the application area, with the idea that what kind of work and area it is used for is the most important thing in utilizing AI. Mr. Mukai of Nextremer says that he is pursuing collaboration between people and AI, AI automatically switches to where people are responding at call centers, etc., or which AI chatbot people use with the touch of a button He said that he is thinking about linking people and AI work well, such as switching.Mr. Kawahara of Retrieva said that elemental technology alone is not enough, but emphasizes how to use AI and how humans can accept AI. What the three speakers shared in common was that technology was there to help people, and it only made sense for people to use it (in practice).

Through this seminar, three speakers who provide solutions centered on natural language processing AI talked about the points derived from actual operation for text data analysis. It is difficult for AI to be fully useful in actual work only with the high accuracy of the AI ​​engine, and there are various issues in utilizing AI, from data management methods to selection of teacher data.This time, the content that the three speakers told from the experience of trying to utilize natural language processing AI for actual operation was a hint for companies that are troubled with the utilization of text data, which is a treasure trove of records. I think that.