Coping with Prospective Memory Failures: An Optimal Reminder System Design

Forgetting is in common in daily life, and 50-80% everyday’s forgetting is due to prospective memory failures, which have significant impacts on our life. More seriously, some of these memory lapses can bring fatal consequences such as forgetting a sleeping infant in the back seat of a car. People tend to use various techniques to improve their prospective memory performance. Setting up a reminder is one of the most important techniques. The existing studies provide evidences in support of using reminders to cope with prospective memory failures. However, people are not satisfied with existing reminders because of their limitations in different aspects including reliability, optimization, and adaption. Through analysing the functions and features of existing reminder systems, this book draft summarizes their advantages and limitations. We are motivated to improve the performance of reminder systems. For the improvements, the relevant theories and mechanisms of prospective memory from psychology must be complied with, incorporated, and applied in this new study. Therefore, prospective memory processes, prospective memory components, and potential factors that influence prospective memory performance were also reviewed in this book draft. Based on the literature review, a new reminder model is proposed, which includes a novel reminder planer, a prospective memory based agent, and a personalized user model. The reminder planer is responsible for determining the optimal reminder plan (including the optimal number of reminders, the optimal reminding schedule and the optimal reminding way). The prospective memory agent is responsible for executing the reminding processes. The personalized user model is proposed to learn from users’ behaviors and preferences based on human-system interactions and is responsible for adapting the reminder plan to meet users’ preferences as much as possible. To realize the functions of different components in our new reminder model, a series of principles and algorithms are presented in the book draft. Our reminder system with the reminder planer, the prospective memory agent, and the personalized user model provides a promising ground for improving the prospective memory performance. The reminder plan, as the product of the reminder planer including the optimal number of reminders, the optimal reminding schedule and the optimal reminding way, overcomes the limitations of current reminder systems in reliability and optimization. Our prospective memory agent supported by psychological models provides a believable diagram of prospective memory processes. It not only demonstrates important components such as planning and monitoring in processes, but also interprets the important stage for prospective memory failures. The execution of reminding integrated into the agent ensures the reminding happens in prospective memory processes to resist prospective memory failures. On the premise of reliability, the personalized user model also acts on the reminder plan to make our system adaptive through learning users’ behaviors from human-system interactions. The future work is required to design and conduct relevant experiments to examine the degree of each factor (the type of ProM task, user’s age, the complexity of the ongoing task, the importance of the ProM task, and the motivation of the ProM task ) and the interactions between these factors. We will apply these results into the system to maximize the probability of remembering to perform the prospective task and minimize the potential annoyance. At the same time, we will improve the methods about learning the users’ behaviors and preferences from human-system interactions.