Increasing the usability of spoken dialogue systems using machine learning : a study of use by older users in retirement homes

Abstract

With a rapidly increasing ageing population, more older people are choosing to be referred to retirement homes. Consequently, this is placing a larger workload on the carers, nurses and allied professionals of said retirement homes. To combat this, many robots and Spoken Dialogue Systems have been designed, however, generally, these are not always used. Studies have shown that older users must make a greater effort to use such systems, when compared to other age groups, leaving them frustrated and less likely to use such systems again. A generic framework for this system usability problem will be designed and a system based on this generic framework will be implemented, focusing on speeding up the chain of commands uttered by the older person, through sequence prediction, over time. Therefore, the core functionalities of the proposed systems will include a voice interface and a machine learning module to learn the sequences that commands are said, over time. It is hoped that by predicting the user’s spoken commands, this system will speed up the process by which an older person reaches their goal, without the aid of a carer, nurse or allied professional. The framework will be tested using elderly users, taking careful note of the time it takes to complete a task, or a set of related tasks, before and after the system learns a sequence of commands. Furthermore, the implemented Spoken Dialogue System will be evaluated for its efficiency and usefulness in retirement homes.