I\'m studying deep learning and recommendation engines. I\'m pretty new to these arguments and my question is:
whatever my starting dataset is (music samples, lyrics,
