To Ask or Not To Ask: Human-in-the-loop Contextual Bandits with Applications in Robot-Assisted Feeding

2024-05-11

arXiv.org

AI Summary

Autonomous feeding strategies in robotic-assisted eating suffer from poor generalization across diverse food types and impose excessive cognitive load on users due to frequent, inflexible queries. Method: This paper proposes a human-robot collaborative contextual bandit framework. It integrates online user studies for behavioral modeling, a synthetic food motion dataset for simulation, and real-world experimental validation. Contribution/Results: We introduce LinUCB-QG—a novel algorithm that jointly models food dynamics and user workload by incorporating query-type semantics and temporal patterns. The framework enables adaptive querying tailored to users with motor impairments. In simulation and a 19-participant user study, our approach significantly outperforms both fully autonomous and always-query baselines, achieving Pareto-optimal trade-offs between task success rate and user burden—particularly for individuals with severe motor impairments.

Technology Category

• Intelligent Robots: Human-Robot Interaction
• Machine Learning: Online Learning & Bandits
• Humans and AI: Intelligent User Interfaces

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Robot-assisted bite acquisition involves picking up food items with varying shapes, compliance, sizes, and textures. Fully autonomous strategies may not generalize efficiently across this diversity. We propose leveraging feedback from the care recipient when encountering novel food items. However, frequent queries impose a workload on the user. We formulate human-in-the-loop bite acquisition within a contextual bandit framework and introduce LinUCB-QG, a method that selectively asks for help using a predictive model of querying workload based on query types and timings. This model is trained on data collected in an online study involving 14 participants with mobility limitations, 3 occupational therapists simulating physical limitations, and 89 participants without limitations. We demonstrate that our method better balances task performance and querying workload compared to autonomous and always-querying baselines and adjusts its querying behavior to account for higher workload in users with mobility limitations. We validate this through experiments in a simulated food dataset and a user study with 19 participants, including one with severe mobility limitations. Please check out our project website at: http://emprise.cs.cornell.edu/hilbiteacquisition/

Problem

• Balancing robot-assisted feeding autonomy
• Minimizing user workload with selective queries
• Adapting to mobility limitations in query strategy

Innovation

• Contextual bandit framework
• LinUCB-QG method
• Workload predictive model