# A Semi-Supervised Deep Learning Approach to Dataset Collection for Query-By-Humming Task

arXiv: https://arxiv.org/abs/2312.01092 · HTML: https://mlsanity.com/paper/2312.01092/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 35 references that are in the map.
- First posted on arXiv 2 Dec 2023.

## Builds on (most cited first, 10 of 10)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Multi-Similarity Loss with General Pair Weighting for Deep Metric Learning](https://mlsanity.com/paper/1904.06627/) (2019)
- [CREPE: A Convolutional Representation for Pitch Estimation](https://mlsanity.com/paper/1802.06182/) (2018)
- [Neural Audio Fingerprint for High-specific Audio Retrieval based on Contrastive Learning](https://mlsanity.com/paper/2010.11910/) (2020)
- [ByteCover: Cover Song Identification via Multi-Loss Training](https://mlsanity.com/paper/2010.14022/) (2020)
- [Learning a Representation for Cover Song Identification Using Convolutional Neural Network](https://mlsanity.com/paper/1911.00334/) (2019)
- [SpliceOut: A Simple and Efficient Audio Augmentation Method](https://mlsanity.com/paper/2110.00046/) (2021)
- [A Prototypical Triplet Loss for Cover Detection](https://mlsanity.com/paper/1910.09862/) (2019)
- [Cover Detection using Dominant Melody Embeddings](https://mlsanity.com/paper/1907.01824/) (2019)
- [Cover Song Identification with Timbral Shape Sequences](https://mlsanity.com/paper/1507.05143/) (2015)

## Abstract (arXiv)

Query-by-Humming (QbH) is a task that involves finding the most relevant song based on a hummed or sung fragment. Despite recent successful commercial solutions, implementing QbH systems remains challenging due to the lack of high-quality datasets for training machine learning models. In this paper, we propose a deep learning data collection technique and introduce Covers and Hummings Aligned Dataset (CHAD), a novel dataset that contains 18 hours of short music fragments, paired with time-aligned hummed versions. To expand our dataset, we employ a semi-supervised model training pipeline that leverages the QbH task as a specialized case of cover song identification (CSI) task. Starting with a model trained on the initial dataset, we iteratively collect groups of fragments of cover versions of the same song and retrain the model on the extended data. Using this pipeline, we collect over 308 hours of additional music fragments, paired with time-aligned cover versions. The final model is successfully applied to the QbH task and achieves competitive results on benchmark datasets. Our study shows that the proposed dataset and training pipeline can effectively facilitate the implementation of QbH systems.