Music recommendations with deep learning
In this thesis we apply deep reinforcement learning to the problem of recom- mending music. A content-based approach is taken, and features from music is extracted with a pretrained deep learning music-tagger. For training, user- interactions are simulated.