Research map / Image generation / Text-to-image diffusion
Melodia: Training-Free Music Editing Guided by Attention Probing in Diffusion Models
Yi Yang, Haowen Li, Tianxiang Li, Boyu Cao, Xiaohan Zhang, Liqun Chen, Qi Liu
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 |
|---|---|---|---|
| replace cross-attention maps (layers 8-14 or 1-16) with source maps (ablation) | AudioLDM 2 editing · Edited temporal structure | structure preservation | Hurt |
| replace self-attention maps in layers 8-14 (ablation) | AudioLDM 2 editing · Edited temporal structure | structure preservation | Improved |
| replace self-attention in all layers 1-16 instead of 8-14 (ablation) | Melodia · MelodiaEdit timbre transfer | LPAPS / ASB: LPAPS 3.49 -> 2.65, ASB 0.68 -> 0.00 | Mixed |
| apply SA replacement in layers 8-14 vs none (ablation) | Melodia · MelodiaEdit timbre transfer | LPAPS: 4.39 -> 3.49 | Improved |
| apply Melodia to Stable Audio Open | Stable Audio Open · MelodiaEdit timbre transfer | LPAPS: 6.47 -> 6.19 | Improved |
| Melodia vs DDPM-Friendly | Melodia · MelodiaEdit | LPAPS: 4.06 -> 3.11 | Improved |
| Melodia vs best baseline on human balance | Melodia · MelodiaEdit | MEB (human): 3.13 -> 3.81 | Improved |
Later papers that compared with this method
1 later accepted paper reports Melodia as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Polyphonia: Zero-Shot Timbre Transfer in Polyphonic Music with Acoustic-Informed Attention Calibration (ICML 2026 · better on 6 of 9 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
- MeLFusion: Synthesizing Music from Image and Language Cues using Diffusion Models (CVPR 2024 · similarity 0.90)
- SongEditor: Adapting Zero-Shot Song Generation Language Model as a Multi-Task Editor (AAAI 2025 · similarity 0.88)
- Prompt-guided Precise Audio Editing with Diffusion Models (ICML 2024 · similarity 0.88)
- MuseControlLite: Multifunctional Music Generation with Lightweight Conditioners (ICML 2025 · similarity 0.87)
- Zero-Shot Unsupervised and Text-Based Audio Editing Using DDPM Inversion (ICML 2024 · similarity 0.87)
- CMI-RewardBench: Evaluating Music Reward Models with Compositional Multimodal Instruction (ICML 2026 · similarity 0.87)
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.
