2023 Fiscal Year Final Research Report
Measuring Group Interaction in Online Discussions and Application to Autonomous Agent Deliberation
| Project/Area Number |
20K11936
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 61030:Intelligent informatics-related
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
Hadfi Rafik 京都大学, 情報学研究科, 特定准教授 (30867495)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
伊藤 孝行 名古屋工業大学, 工学(系)研究科(研究院), 教授 (50333555)
|
|---|
| Project Period (FY) |
2020-04-01 – 2024-03-31
|
| Keywords | Agents / Conversational AI / Decision-making / Interaction / Deliberation / Similarity Metrics / Online Discussion / Entropy Methods |
|---|
| Outline of Final Research Achievements |
The research achievements involve the discovery various interaction patterns in group discussions involving both humans and agents. These findings have been disseminated through international journals (E.g., Scientific Reports, Social Network Analysis and Mining), as well as national and international conferences, workshops, and tutorials (E.g., AAMAS, PRICAI, JSAI, IEEE ICA, etc.). One significant achievement includes the development of an evaluation method published in the journal Social Network Analysis and Mining. The PI explored whether the structural complexity of online discussions could predict consensus. This method combines metrics from well-known readability tests with an entropy-based complexity metric applied to the tree structures of Reddit discussions. The findings indicate that the proposed metric effectively predicts consensus readability based solely on the complexity of discourse structure. These achievements are now being extended in other ongoing projects.
|
| Free Research Field |
Artificial intelligence
|
| Academic Significance and Societal Importance of the Research Achievements |
The research leads to agents that can assist humans in addressing social problems. E.g., the study in "Scientific Reports", demonstrated that AI can enhance the social presence of women. Another application is the "Computational Social Choice Competition", which advances the research on democracy.
|
