These notes are a summary of concepts presented in “Interaction-Shaping Robotics: Robots That Influence Interactions between Other Agents.”
Sarah Gillet, Marynel Vázquez, Sean Andrist, Iolanda Leite, and Sarah Sebo. 2024. Interaction-Shaping Robotics: Robots That Influence Interactions between Other Agents. ACM Trans. Hum.-Robot Interact. 13, 1,
Article 12 (March 2024), 23 pages. https://doi.org/10.1145/3643803
- Interaction-Shaping Robotics
- A subfield of HRI focused on robots influencing interactions between two or more agents
- Key aspects
- Role of robot
- Robot-shaping outcomes
- Form of influence
- Type of communication
- Influence timeline
- Effects of Robot Behavior
- Direct Reciprocal Effect
- Traditional HRI scenarios where the robot influences human behaviors back toward itself
- Indirect Effect
- Unique to Interaction-Shaping Robotics, where the robot shapes the interaction or perception between agents in a group
- Direct Reciprocal Effect
- Key Factors in ISR
- Role of Robot
- Guiding Facilitator: Actively mediates interaction
- Peripheral Facilitator: Active but not directly involved
- Peer Group Member: Participates as a peer
- Specialized Group Member: Assumes a specific role
- Robot-Shaping Outcomes
- Cognitive: Influences attitudes, evaluations, and intentions
- Behavioral: Affects actions like speaking or spatial positioning
- Form of Robot Influence
- Explicit: Direct communication (e.g., conflict resolution)
- Implicit: Subtle cues shaping interaction (e.g., gestures)
- Type of Robot Communication
- Verbal: Natural language
- Non-Verbal: Gestures, gaze, movements
- Timeline of Influence
- Immediate: Real-time interaction shaping
- Long-Lasting: Post-interaction effects
- Role of Robot
- Communication Capabilities of Robots
- Beyond human abilities: Use of lights, sounds, and movements unique to robots
- Example structures
- Single robot influencing human-human interaction
- Single robot in human-robot interaction
- Multiple robots influencing group interactions
- Ethical Considerations in Interaction-Shaping Robotics
- Reducing Risks
- Dependency: Robots aim to sustainably improve interactions, eventually becoming obsolete
- Deception: Balancing awareness and ethical use of influence, following IEEE guidelines
- Bias Mitigation
- Risks from societal and dataset biases, particularly in machine learning-based behaviors
- Awareness Spectrum
- Varies from fully aware to unaware, influenced by non-verbal communication or robot roles
- Reducing Risks
- Social Dynamics and Group Influence
- Leveraging and inducing group phenomena like social cohesion
- Importance of modeling relationships and group factors in human-robot interaction
- Bridging Social Signal Processing (SSP) and ISR for improved perception and interaction modeling
- Technological Pathways
- Graph Abstractions
- Encoding interactants and relationships as nodes and edges for structured data representation
- Learning Algorithms
- Reinforcement Learning and Imitation Learning for mapping group states to effective behaviors
- Early exploration: Balancing participation in conversations.
- Graph Abstractions
- Future Directions
- Enhancing Interaction-shaping robotics perception and control loops
- Developing algorithms that address ethical, social, and technical challenges
- Advancing methodologies to sustain and enrich human interactions and relationships