An automated choreographer is proposed to choose musically expressive dance moves, with a ran-domized move selection to preclude prediction by judges or audiences. Dance moves and music bars are scored on emotional attribute scales. At each bar, an optimization model ranks the move alternatives on their musical appropriateness, and a move is selected from a distribution biased towards the highest ranked moves. This biased distribution is discovered from performer-created dance routines, and the optimization model is validated by consistent bias across routines. The role of surprise in music and dance is also examined. The automated choreographer will find application in any setting in which optimizing some objective must be balanced with retaining the element of surprise. © 2004 IEEE.
|Original language||English (US)|
|Title of host publication||Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics|
|Number of pages||6|
|State||Published - Dec 1 2004|