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 |
Rafik Hadfi 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
|
| Project Status |
Completed (Fiscal Year 2023)
|
| Budget Amount *help |
¥3,770,000 (Direct Cost: ¥2,900,000、Indirect Cost: ¥870,000)
Fiscal Year 2022: ¥390,000 (Direct Cost: ¥300,000、Indirect Cost: ¥90,000)
Fiscal Year 2021: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2020: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
|
|---|
| Keywords | Agents / Conversational AI / NLP / Deliberation / Interaction / Decision-making / Interdependence / Online Discussion / Online Debates / Agent Deliberation / Predictive Deliberation / Conversational Agents / Augmented Democracy / Mutual Information / Natural Language / Group Interaction / Similarity Metrics / Collaborative Editing / Time Series Analysis / Automatic Deliberation / Entropy Methods / Artificial Agent / Information Theory / Artificial Intelligence |
|---|
| Outline of Research at the Start |
In this project, we propose to study the interactions between humans and artificial agents that maximize collective intelligence. Understanding the dynamics behind symbiotic interactions in online discussions is a viable way to foster intelligent deliberation and build smarter deliberative agents.
|
| Outline of Annual Research Achievements |
This final phase's achievement is applying previous metrics to practical cases of humans interacting with agents in a computational social study. The study was published in Nature's Scientific Reports and focuses on citizens participating in online debates, demonstrating that conversational agents can influence discussion dynamics, enhancing participation and reducing inhibition. The second key result builds on interaction metrics to propose a study published in the Journal of Social Network Analysis and Mining. I explored whether the structural complexity of online discussions can predict consensus readability without linguistic semantics. The findings indicate that entropy-based metrics effectively predict consensus readability based on the complexity of the discourse tree. These findings contribute to the symbiotic interactions between humans and intelligent deliberative agents in online discussion platforms.
|
Report
(4 results)
Research Products
(21 results)
