Representation Learning For Music And Audio Intelligence