Computational discovery of personal traits from social multimedia

Hundreds of millions of people leave digital footprints, including textual posts and photos, on public on social media and social networking sites. Here we present our work on using these digital footprints-social multimedia content-to derive four types of basic personal traits for individuals. Moreover, we show how these basic traits can be used in combinations to help assess composite traits in two cases: trust modeling and resilience modeling. We share our preliminary results and discuss future research directions.