ACT@HOME: AFFECTIVE AND COGNITIVE TECHNOLOGIES FOR ASSISTANCE IN THE HOME

Abstract

Our overall aim is to develop an emotionally intelligent cognitive assistant (ICA) to engage and help older adults with Alzheimer's disease (AD) to complete activities of daily living (ADL) more independently. The system will take emotional factors into account, such as how the person feels about who they are (their emotional identity) and how they feel about the system (their sentiments about technologies, for example). In developing this system, the project will study emotional identity and behavior in older adults with AD, leading to a computational characterization of these aspects. To accomplish this, we first carry out semi-structured qualitative interviews with elderly nursing home residents suffering from mild AD in order to extract their different affective identities, personalities and backgrounds. For this, a specific new interview tool was designed based on the principles of Affect Control Theory (ACT), a socio-cultural theory of affective interactions between persons. To assess identity, the questionnaire asked questions about life domains (family and origin, occupation/vocation, personal history such as immigration or military service, and relationships associated with strong feelings such as traumatic or deeply romantic ones). In a next step, all interviews are transcribed and analyzed in order to program the system to learn the ‘numerical terms’ for each affective identity of a person during an interaction extracted from these interviews, and will then tailor prompts to specific individual’s needs’ in a way that ensures smoother and more effective uptake and response. Preliminary results of the first analysis of the interviews show that we can distinguish clearly between certain affective identities, for instance ‘the carrying wife’, the ‘authoritarian father’ etc. in this population and thus, define their resulting preferences in a specific prompting style provided by the ICA. An extensive overview of the analyses results will be presented in depth at the conference and build the basis for the further development of the novel better to its end-user adapted ICA. The findings from this research will continue to progress towards achieving the long-term goal of technology to support older adults with AD. This will be accomplished through continuing with the expansion of the system to other ADL. Our results will have extrinsic merit as they will have a significant impact not only on the AD field and smart home systems, but also in the affective computing field. We will attempt to produce outcomes that will answer questions related to emotional interactions between humans and machines.