Research map / Label-efficient and robust learning / Semi-supervised learning
Every Little Bit Helps: Exploring Better Utilization of Unlabeled Data for Semi-supervised Singing Melody Extraction Using Multi-bands Diffusion Model
Shuai Yu, Xiaoliang He, Kangjie Dong, Yi Yu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add diffusion multi-band augmentation (DMA) (ablation) | ELH-SME · ADC2004 | OA (%): 80.4 -> 84.4 | Improved |
| add diffusion multi-band augmentation (DMA) (ablation) | ELH-SME · MIREX 05 | OA (%): 83.1 -> 87.6 | Improved |
| add global-class confidence (GCC) module (ablation) | ELH-SME · ADC2004 | OA (%): 82.1 -> 84.4 | Improved |
| add global-class confidence (GCC) module (ablation) | ELH-SME · MIREX 05 | OA (%): 85.9 -> 87.6 | Improved |
| add channel cross-attention (CCA) module (ablation) | ELH-SME · ADC2004 | OA (%): 82.7 -> 84.4 | Improved |
| add channel cross-attention (CCA) module (ablation) | ELH-SME · MIREX 05 | OA (%): 85.2 -> 87.6 | Improved |
| use multi-band DMA instead of plain DDPM on the spectrogram (ablation) | ELH-SME · ADC2004 | OA (%): 81.2 -> 84.4 | Improved |
| use multi-band DMA instead of plain DDPM on the spectrogram (ablation) | ELH-SME · MIREX 05 | OA (%): 85.9 -> 87.6 | Improved |
| use multi-band DMA instead of low-frequency-only augmentation (ablation) | ELH-SME · ADC2004 | OA (%): 81.1 -> 84.4 | Improved |
| use multi-band DMA instead of high-frequency-only augmentation (ablation) | ELH-SME · ADC2004 | OA (%): 80.2 -> 84.4 | Improved |
| use ELH-SME instead of MCSSME | ELH-SME · ADC2004 | OA (%): 76.7 -> 84.4 | Improved |
| use ELH-SME instead of MCSSME | ELH-SME · MIREX 05 | OA (%): 86.6 -> 87.6 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MCSSME — MCSSME: Multi-Task Contrastive Learning for Semi-supervised Singing Melody Extraction from Polyphonic Music (AAAI 2024) · better on 8 of 8 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MCSSME: Multi-Task Contrastive Learning for Semi-supervised Singing Melody Extraction from Polyphonic Music (AAAI 2024 · similarity 0.91)
- MGE-LDM: Joint Latent Diffusion for Simultaneous Music Generation and Source Extraction (NeurIPS 2025 · similarity 0.85)
- GTSinger: A Global Multi-Technique Singing Corpus with Realistic Music Scores for All Singing Tasks (NeurIPS 2024 · similarity 0.84)
- Weakly-supervised Audio Separation via Bi-modal Semantic Similarity (ICLR 2024 · similarity 0.83)
- MERT: Acoustic Music Understanding Model with Large-Scale Self-supervised Training (ICLR 2024 · similarity 0.82)
- Synthio: Augmenting Small-Scale Audio Classification Datasets with Synthetic Data (ICLR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
