Designing an AI Health Coach and Studying its Utility in Promoting Regular Aerobic Exercise (1910.04836v2)

Abstract: Our research aims to develop interactive, social agents that can coach people to learn new tasks, skills, and habits. In this paper, we focus on coaching sedentary, overweight individuals (i.e., trainees) to exercise regularly. We employ adaptive goal setting in which the intelligent health coach generates, tracks, and revises personalized exercise goals for a trainee. The goals become incrementally more difficult as the trainee progresses through the training program. Our approach is model-based - the coach maintains a parameterized model of the trainee's aerobic capability that drives its expectation of the trainee's performance. The model is continually revised based on trainee-coach interactions. The coach is embodied in a smartphone application, NutriWalking, which serves as a medium for coach-trainee interaction. We adopt a task-centric evaluation approach for studying the utility of the proposed algorithm in promoting regular aerobic exercise. We show that our approach can adapt the trainee program not only to several trainees with different capabilities, but also to how a trainee's capability improves as they begin to exercise more. Experts rate the goals selected by the coach better than other plausible goals, demonstrating that our approach is consistent with clinical recommendations. Further, in a 6-week observational study with sedentary participants, we show that the proposed approach helps increase exercise volume performed each week.

• The paper presents an AI-driven health coach that adapts daily and weekly aerobic exercise goals based on personalized performance data.
• Key methodology includes a staircase model for aerobic capability and adaptive goal-setting validated by expert feedback and simulations.
• The 6-week evaluation demonstrates increased exercise engagement among sedentary participants with a mean usability score of 75/100.

Designing an AI Health Coach and Studying its Utility in Promoting Regular Aerobic Exercise

Introduction

The paper investigates the development of an AI-based health coach, particularly targeting sedentary and overweight individuals to promote aerobic exercise as a healthy lifestyle. The intelligent agent is implemented through a mobile application, NutriWalking, which applies model-based adaptive goal setting. This agent tracks and revises personalized exercise goals for a trainee, with difficulty adjustments as trainees progress.

Methodology

Adaptive Goal Setting

The coach employs a structured computational approach to goal setting, which is divided into weekly and daily scheduling. It personalizes coaching by maintaining a parameterized model of a trainee's aerobic capability, adjusting goals based on observed performance.

The coaching strategy follows the principles of setting challenging yet attainable, specific, and proximal goals.

Figure 1: Various human-agent interactions NutriWalking - (top) initial goal setting, (bottom left) main dashboard, (bottom center) daily goal reporting, (bottom right) daily question answering.

Trainee Aerobic Capability Model

The paper presents a staircase model for aerobic capability growth, parameterized to represent variations in trainee profiles and capabilities. It uses model revisions to adapt the program as trainees progress or show variance in goal compliance.

Figure 2: An adaptable model for aerobic capability. Blue line (CIRCLE) represents the staircase model for growth in capability and green (blacksquare), yellow (times), and red (blacktriangle) its revisions.

Evaluation

Adaptivity, Suitability, and Usability

The AI coach's adaptability was assessed through simulations of varying trainee profiles, capable of temporal and personal goal adaptations.

Empirical evaluations with expert physical therapists validated the suitability of the generated exercise goals, finding them to be safely attainable.

Usability studies exhibited positive feedback among individuals with type 2 diabetes or depression, who provided a mean System Usability Scale score of 75/100, indicating good acceptance.

Observed Impact Over 6 Weeks

Through a 6-week paper with sedentary participants, the adaptive coach demonstrated a positive effect on exercise volume. It suggested incremental goal adjustments and revealed participant engagement maintained throughout the program duration.

Figure 3: Weekly goals, exercise volume performed (xvolume), and rate of perceived exertion (RPE) of two participants for the length of the paper.

Limitations and Discussion

The research notes several limitations including reliance on self-assessment, the challenge of overly optimistic self-reporting skewing modeling accuracy, and platform constraints limited to iOS. Real-world applications presented some degree of trainee overestimation, necessitating a more accurate benchmarking process for initial aerobic capability.

The research advocates for broader inclusion of behavioral elements such as motivation levels and mood in future iterations, aiming to refine goal predictions and enhance program compliance and success.

Conclusion

The results of the paper underscore the potential for AI-driven solutions to support learning and adaptation in promoting healthier lifestyles via regular aerobic exercise. The NutriWalking intelligent coach proved to augment exercise through tailored goals for different individuals and contexts. Through continuous improvements and a comprehensive approach incorporating more interactive metrics, this AI methodology can foster significant societal health welfare advancements.