Using an affective human-like association architecture for life-logs
Jana Machajdik, Lukas Kavicky · International Conference on Systems, Signals and Image Processing · 2012
Diaries - and now lifelogs - offer ways to store old memories, review them from time to time, and reflect about the past. Self-reflection is a crucial factor in personal growth. However, to make this process effective the data from lifelogs has to be structured in a way that is meaningful for humans. In this paper we presented the design of an affective human-like associations architecture that creates models inspired by the way humans create associations in their unconscious. This is done by interconnecting the multimodal signals collected by lifelogging sensors in a flexible way, using similarity associations and affective links in combination with an adaptation of the spreading activation theory of memory. The approach enables the user to use these affective associations to browse his or her lifelog data in a way that encourages self-reflection by suggesting similar situations.