Transfer learning for music classification and regression tasks
This work addresses music analysis tasks for researchers and practitioners, but it is incremental as it applies an existing transfer learning method to a specific domain.
The authors tackled music classification and regression tasks by using a pre-trained convolutional network feature as a general-purpose music representation, achieving performance improvements over baseline MFCC features and previous approaches in all considered tasks.
In this paper, we present a transfer learning approach for music classification and regression tasks. We propose to use a pre-trained convnet feature, a concatenated feature vector using the activations of feature maps of multiple layers in a trained convolutional network. We show how this convnet feature can serve as general-purpose music representation. In the experiments, a convnet is trained for music tagging and then transferred to other music-related classification and regression tasks. The convnet feature outperforms the baseline MFCC feature in all the considered tasks and several previous approaches that are aggregating MFCCs as well as low- and high-level music features.