IBM Corporation, 2017 IBM Faculty Award(知能システム学専攻 長尾確 教授)

受賞日: 平成29年7月6日
受賞者: 長尾 確 教授(情報学研究科 知能システム学専攻)
受賞概要: “Meeting Analytics: Creative Activity Support by Knowledge Discovery from Discussions” の研究プロジェクトに対する受賞


Our research aims to deeply analyze face-to-face meetings and support human creative activities by appropriately feeding back knowledge discovered from the meetings.
For that purpose, we are developing a system called a meeting recorder.
The meeting recorder consists of a 360 degree panoramic video camera that records meeting audio-visual scenes, a tablet application that allows users to browse meeting materials and to add them some notes with a pen, speech recognition that identifies speakers and transcribes speech contents of all meeting participants, and a minute server that integrates all meeting related information and creates minutes of the meeting.
On the minute server, machine learning is used to estimate statements that change topics in discussion.
As a result, the minutes are divided for each topic, and priorities are determined in consideration of the number of statements and speakers in the topic.
If the discussions can be further structured, more sophisticated processing such as discovery of creativity and acceleration of innovation can be realized.
In order to structure them, it is necessary to gather a lot of data.
We are planning to record and analyze actual meetings in large quantities at Nagoya University and IBM Tokyo Research Laboratory.