Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Moderating Group Conversation Dynamics with Social Robots

Lucrezia Grassi, Carmine Tommaso Recchiuto, Antonio Sgorbissa

TL;DR

This work tackles the problem of moderating group conversations with social robots in multi-user settings. It presents CAIR, a cloud-based framework with an ontology-driven Dialogue/Plan management stack and a moving-window mechanism to track participation, enabling two policy families—Balancing and Community—with hard and soft variants. Empirical results from 75 four-person groups (300 participants) show that Balancing policies reduce disparities in speaking time and word usage, while Community policies curb subgroup formation, with strong statistical significance ($p<0.01$). The findings highlight practical implications for inclusive, coordinated multi-user interactions in domains like education, healthcare, and entertainment, while respecting privacy through data deletion after experiments.

Abstract

This research investigates the impact of social robot participation in group conversations and assesses the effectiveness of various addressing policies. The study involved 300 participants, divided into groups of four, interacting with a humanoid robot serving as the moderator. The robot utilized conversation data to determine the most appropriate speaker to address. The findings indicate that the robot's addressing policy significantly influenced conversation dynamics, resulting in more balanced attention to each participant and a reduction in subgroup formation.

Moderating Group Conversation Dynamics with Social Robots

TL;DR

This work tackles the problem of moderating group conversations with social robots in multi-user settings. It presents CAIR, a cloud-based framework with an ontology-driven Dialogue/Plan management stack and a moving-window mechanism to track participation, enabling two policy families—Balancing and Community—with hard and soft variants. Empirical results from 75 four-person groups (300 participants) show that Balancing policies reduce disparities in speaking time and word usage, while Community policies curb subgroup formation, with strong statistical significance (). The findings highlight practical implications for inclusive, coordinated multi-user interactions in domains like education, healthcare, and entertainment, while respecting privacy through data deletion after experiments.

Abstract

This research investigates the impact of social robot participation in group conversations and assesses the effectiveness of various addressing policies. The study involved 300 participants, divided into groups of four, interacting with a humanoid robot serving as the moderator. The robot utilized conversation data to determine the most appropriate speaker to address. The findings indicate that the robot's addressing policy significantly influenced conversation dynamics, resulting in more balanced attention to each participant and a reduction in subgroup formation.
Paper Structure (8 sections, 2 equations, 4 figures, 1 table)

This paper contains 8 sections, 2 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methodology
  3. System Architecture
  4. Control Policies
  5. Balancing Policy
  6. Community Policy
  7. Experiments and Results
  8. Conclusion

Figures (4)

  • Figure 1: Multi-party interaction between the humanoid robot Pepper and a group of four students of the “Parini Merello" middle school in Genoa, during an experiment performed for this study.
  • Figure 2: CAIR system architecture
  • Figure 3: Comparison of the speaking times of the participants when interacting with the robot using the Neutral policy and the Balancing policy in its soft version.
  • Figure 4: Comparison of the number of communities at each conversation turn when applying the Neutral policy and the Community policy in its hard version.