A Transformer Based Pitch Sequence Autoencoder with MIDI Augmentation
This work addresses the need for automated detection of AI-generated music, which is incremental as it builds on existing transformer models with specific optimizations.
The paper tackles the problem of classifying whether a single-track music excerpt is composed by humans or AI, using a masked language model based on ALBERT with reduced parameters, data augmentation, and a refined loss function. The result shows the model ranks 3rd out of 7 teams in a data challenge.
Despite recent achievements of deep learning automatic music generation algorithms, few approaches have been proposed to evaluate whether a single-track music excerpt is composed by automatons or Homo sapiens. To tackle this problem, we apply a masked language model based on ALBERT for composers classification. The aim is to obtain a model that can suggest the probability a MIDI clip might be composed condition on the auto-generation hypothesis, and which is trained with only AI-composed single-track MIDI. In this paper, the amount of parameters is reduced, two methods on data augmentation are proposed as well as a refined loss function to prevent overfitting. The experiment results show our model ranks $3^{rd}$ in all the $7$ teams in the data challenge in CSMT(2020). Furthermore, this inspiring method could be spread to other music information retrieval tasks that are based on a small dataset.