Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring Causality for HRI: A Case Study on Robotic Mental Well-being Coaching

Micol Spitale, Srikar Babu, Serhan Cakmak, Jiaee Cheong, Hatice Gunes

TL;DR

The paper investigates how causality can enhance adaptive robotic coaching for mental well-being in real-world HRI. By combining a macro-level structural equation modeling approach with a micro-level fast causal inference analysis on a four-week QTrobot dataset, it links multimodal cues (facial actions and speech features) to wellbeing and perceived robot alliance. Key findings show specific audio and facial cues correlate with ROSAS, PANAS, and WAI, and that latent confounders shape turn-level interactions; the work demonstrates the value of causality for explainable, adaptable HRI. Limitations include reliance on observational data and lack of counterfactual estimation, with future work exploring additional causal estimators and broader HRI contexts, while preserving participant privacy.

Abstract

One of the primary goals of Human-Robot Interaction (HRI) research is to develop robots that can interpret human behavior and adapt their responses accordingly. Adaptive learning models, such as continual and reinforcement learning, play a crucial role in improving robots' ability to interact effectively in real-world settings. However, these models face significant challenges due to the limited availability of real-world data, particularly in sensitive domains like healthcare and well-being. This data scarcity can hinder a robot's ability to adapt to new situations. To address these challenges, causality provides a structured framework for understanding and modeling the underlying relationships between actions, events, and outcomes. By moving beyond mere pattern recognition, causality enables robots to make more explainable and generalizable decisions. This paper presents an exploratory causality-based analysis through a case study of an adaptive robotic coach delivering positive psychology exercises over four weeks in a workplace setting. The robotic coach autonomously adapts to multimodal human behaviors, such as facial valence and speech duration. By conducting both macro- and micro-level causal analyses, this study aims to gain deeper insights into how adaptability can enhance well-being during interactions. Ultimately, this research seeks to advance our understanding of how causality can help overcome challenges in HRI, particularly in real-world applications.

Exploring Causality for HRI: A Case Study on Robotic Mental Well-being Coaching

TL;DR

The paper investigates how causality can enhance adaptive robotic coaching for mental well-being in real-world HRI. By combining a macro-level structural equation modeling approach with a micro-level fast causal inference analysis on a four-week QTrobot dataset, it links multimodal cues (facial actions and speech features) to wellbeing and perceived robot alliance. Key findings show specific audio and facial cues correlate with ROSAS, PANAS, and WAI, and that latent confounders shape turn-level interactions; the work demonstrates the value of causality for explainable, adaptable HRI. Limitations include reliance on observational data and lack of counterfactual estimation, with future work exploring additional causal estimators and broader HRI contexts, while preserving participant privacy.

Abstract

One of the primary goals of Human-Robot Interaction (HRI) research is to develop robots that can interpret human behavior and adapt their responses accordingly. Adaptive learning models, such as continual and reinforcement learning, play a crucial role in improving robots' ability to interact effectively in real-world settings. However, these models face significant challenges due to the limited availability of real-world data, particularly in sensitive domains like healthcare and well-being. This data scarcity can hinder a robot's ability to adapt to new situations. To address these challenges, causality provides a structured framework for understanding and modeling the underlying relationships between actions, events, and outcomes. By moving beyond mere pattern recognition, causality enables robots to make more explainable and generalizable decisions. This paper presents an exploratory causality-based analysis through a case study of an adaptive robotic coach delivering positive psychology exercises over four weeks in a workplace setting. The robotic coach autonomously adapts to multimodal human behaviors, such as facial valence and speech duration. By conducting both macro- and micro-level causal analyses, this study aims to gain deeper insights into how adaptability can enhance well-being during interactions. Ultimately, this research seeks to advance our understanding of how causality can help overcome challenges in HRI, particularly in real-world applications.

Paper Structure

This paper contains 15 sections, 2 equations, 3 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Causality in HRI
  4. Robotic mental well-being coaching
  5. Robotic Well-being Coaching Interactions Dataset
  6. Methodology
  7. Research questions
  8. Macro-level Analysis: Structural Equation Modeling
  9. Micro-level Analysis: Fast Causal Inference (FCI)
  10. Experiments and Results
  11. Macro Analysis Results
  12. Audio Features
  13. Facial Features
  14. Micro Analysis Results
  15. Summary & Limitation

Figures (3)

  • Figure 1: Macro (a, top) and micro (b, bottom) analysis representations.
  • Figure 2: A sample SEM model showing the relation between ROSAS metric and features - spectal flux and Loudness.
  • Figure 3: The resulting causal graph obtained using the FCI algorithm. Legend: (1) $X \to Y$, indicating X causes Y; (2) $X \leftrightarrow Y$, suggesting unmeasured confounders affecting both X and Y; (3) X o$\rightarrow$ Y, meaning either X causes Y or there are unmeasured confounders affecting both; and (4) X o-o Y, representing several possibilities, such as X causing Y, Y causing X, or the influence of latent confounders from both.