Research map / Label-efficient and robust learning / Self-supervised and contrastive
MAPSS: Manifold-based Assessment of Perceptual Source Separation
Amir Ivry, Samuele Cornell, Shinji Watanabe
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 |
|---|---|---|---|
| use self-supervised (wav2vec 2.0) encoding instead of raw waveform in PS (ablation) | PS · SEBASS English | SRCC: 73.42% -> 84.12% | Improved |
| use self-supervised encoding instead of raw waveform in PM (ablation) | PM · SEBASS English | SRCC: 69.30% -> 84.69% | Improved |
| PS/PM versus PESQ on English SRCC | PS · SEBASS English | SRCC: 85.56% (PESQ) vs 84.12% (PS) | Hurt |
| PS/PM versus SDR-based metrics on Spanish SRCC | PS · SEBASS Spanish | SRCC: 84.29% (SDR) vs 82.33% (PS) | Hurt |
| PS/PM versus 14 baseline measures on music | PS/PM · SEBASS Music (No Drums) | SRCC: 75.64% (STOI) -> 87.23% (PS) | Improved |
| diffusion maps with PM versus cosine-similarity SpeechBERTScore | PM · SEBASS English | PCC: 67.44% -> 86.36% | Improved |
| use wav2vec 2.0 Large instead of Base for PM (ablation) | PM · SEBASS English | PCC: 80.07% -> 86.36% | Improved |
| use different transformer layers 1-3 (ablation) | PS/PM · SEBASS speech | SRCC/PCC | No clear effect |
| use PS instead of PM for confidence interval width | PS · SEBASS English | SRCC 95% CI: 7.23% (PM) vs 30.03% (PS) | Hurt |
Closest papers by meaning
- Seeing Speech and Sound: Distinguishing and Locating Audio Sources in Visual Scenes (CVPR 2025 · similarity 0.86)
- Seeing Sound, Hearing Sight: Uncovering Modality Bias and Conflict of AI models in Sound Localization (NeurIPS 2025 · similarity 0.84)
- ADIFF: Explaining audio difference using natural language (ICLR 2025 · similarity 0.84)
- Weakly-supervised Audio Separation via Bi-modal Semantic Similarity (ICLR 2024 · similarity 0.84)
- Massive Sound Embedding Benchmark (MSEB) (NeurIPS 2025 · similarity 0.84)
- Multi-Source Diffusion Models for Simultaneous Music Generation and Separation (ICLR 2024 · similarity 0.84)
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.
