Concurrent Training Of A Control Policy