Reward Design for an Online Reinforcement Learning Algorithm Supporting Oral Self-Care

While dental disease is largely preventable, professional advice on optimal oral hygiene practices often forgotten or abandoned by patients. Therefore patients may benefit from timely and personalized encouragement to engage in self-care behaviors. In this paper, we develop an online reinforcement learning (RL) algorithm for use optimizing the delivery of mobile-based prompts encourage One main challenges developing such ensuring that considers impact current actions effectiveness future (i.e., delayed effects), especially when has been designed run stably autonomously a constrained, real-world setting characterized highly noisy, sparse data. We address challenge designing quality reward maximizes desired health outcome high-quality brushing) while minimizing user burden. also highlight procedure hyperparameters building simulation environment test bed evaluating candidates using bed. The RL discussed paper will be deployed Oralytics. To best our knowledge, Oralytics first mobile study utilizing prevent motivational messages supporting