Footprints on Silicon: Explorations in Gathering Autobiographical Content

As we interact with the world, we leave behind digital trails in the form of emails, blogs, tweets and posts, which serve as a rich source of data for generating our individual life stories, or autobiographies. Central to addressing the problem is the ability to discriminate content that is of autobiographical value from the rest. The features required for this classification task need to be discovered from the unstructured data, metadata, sentiments, properties of the social network and temporal properties of the interactions. In this paper we identify several dimensions of this problem, present some preliminary results on our explorations, and identify interesting research problems for the future.