MARE: Recognition and Discovery Techniques for Giving Awarenesss

2023 Fiscal Year Final Research Report

• Project/Area Number: 19H04128
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Review Section: Basic Section 61010:Perceptual information processing-related
• Research Institution: Hokkaido University
• Principal Investigator: Kudo Mineichi 北海道大学, 情報科学研究院, 教授 (60205101)
• Co-Investigator(Kenkyū-buntansha): 今井 英幸 北海道大学, 情報科学研究院, 教授 (10213216) ; 中村 篤祥 北海道大学, 情報科学研究院, 教授 (50344487)
• Project Period (FY): 2019-04-01 – 2024-03-31
• Keywords: 不頻出事象 / マルチラベル分類 / 希少疾患の発見 / 高齢者の異変検知

Outline of Final Research Achievements

On the standing point that rare events, such as rare diseases, anomaly of elderly living alone and interesting but unknown area, are more important than usual events, we have developed techniques for discovering rare events and classifying them. The reasons of occurrence of rare events were revealed to be "new" and devised from the necessity of finer labels. Accordingly, on the basis of these reasons, we developed some classifiers, but only a small amount of improvement was achieved by them.
As an application, on discovery of rate diseases, we have shown that visualization is useful for distinguishing rate events from ordinary events. For detecting anomalies of elderly living alone, we have developed a behavior simulator in a virtual smart home and detection algorithms of anomaly detection.

Free Research Field

パターン認識

Academic Significance and Societal Importance of the Research Achievements

不頻出事象はこの世界にあまねく存在するものの、その希少性により扱いが難しい。本研究はそれらの分類についての方法論を深化させた。特に、専門家でも難しい希少疾患の発見の一助になる方法を提示できたこと、スマートホームを利用して独居高齢者の異変を検知する試みの基本的有効性を示せたことの意義は大きい。