Improving BERT for Symbolic Music Understanding Using Token Denoising and Pianoroll Prediction
This work addresses symbolic music understanding for researchers and practitioners, representing an incremental improvement with novel pre-training objectives.
The paper tackled symbolic music understanding by proposing a BERT-like model with token denoising and pianoroll prediction pre-training objectives, achieving competitive performance across 12 downstream tasks such as chord estimation and genre classification.
We propose a pre-trained BERT-like model for symbolic music understanding that achieves competitive performance across a wide range of downstream tasks. To achieve this target, we design two novel pre-training objectives, namely token correction and pianoroll prediction. First, we sample a portion of note tokens and corrupt them with a limited amount of noise, and then train the model to denoise the corrupted tokens; second, we also train the model to predict bar-level and local pianoroll-derived representations from the corrupted note tokens. We argue that these objectives guide the model to better learn specific musical knowledge such as pitch intervals. For evaluation, we propose a benchmark that incorporates 12 downstream tasks ranging from chord estimation to symbolic genre classification. Results confirm the effectiveness of the proposed pre-training objectives on downstream tasks.