Abstract
This paper asks to what extent social interaction influences one's behavior.We study this in the setting of two dancers dancing as a couple. We firstconsider a baseline in which we predict a dancer's future moves conditionedonly on their past motion without regard to their partner. We then investigatethe advantage of taking social information into account by conditioning also onthe motion of their dancing partner. We focus our analysis on Swing, a dancegenre with tight physical coupling for which we present an in-the-wild videodataset. We demonstrate that single-person future motion prediction in thiscontext is challenging. Instead, we observe that prediction greatly benefitsfrom considering the interaction partners' behavior, resulting in surprisinglycompelling couple dance synthesis results (see supp. video). Our contributionsare a demonstration of the advantages of socially conditioned future motionprediction and an in-the-wild, couple dance video dataset to enable futureresearch in this direction. Video results are available on the project website:https://von31.github.io/synNsync