AI Paper Atlas

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

AAAI 2026 · Label-efficient and robust learning › Semi-supervised learning · approach: Holistic semi-supervised pipelines · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
add diffusion multi-band augmentation (DMA) (ablation)ELH-SME · ADC2004OA (%): 80.4 -> 84.4Improved
add diffusion multi-band augmentation (DMA) (ablation)ELH-SME · MIREX 05OA (%): 83.1 -> 87.6Improved
add global-class confidence (GCC) module (ablation)ELH-SME · ADC2004OA (%): 82.1 -> 84.4Improved
add global-class confidence (GCC) module (ablation)ELH-SME · MIREX 05OA (%): 85.9 -> 87.6Improved
add channel cross-attention (CCA) module (ablation)ELH-SME · ADC2004OA (%): 82.7 -> 84.4Improved
add channel cross-attention (CCA) module (ablation)ELH-SME · MIREX 05OA (%): 85.2 -> 87.6Improved
use multi-band DMA instead of plain DDPM on the spectrogram (ablation)ELH-SME · ADC2004OA (%): 81.2 -> 84.4Improved
use multi-band DMA instead of plain DDPM on the spectrogram (ablation)ELH-SME · MIREX 05OA (%): 85.9 -> 87.6Improved
use multi-band DMA instead of low-frequency-only augmentation (ablation)ELH-SME · ADC2004OA (%): 81.1 -> 84.4Improved
use multi-band DMA instead of high-frequency-only augmentation (ablation)ELH-SME · ADC2004OA (%): 80.2 -> 84.4Improved
use ELH-SME instead of MCSSMEELH-SME · ADC2004OA (%): 76.7 -> 84.4Improved
use ELH-SME instead of MCSSMEELH-SME · MIREX 05OA (%): 86.6 -> 87.6Improved

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.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy