AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

PADS-TAL: Padding-Annealed Diffusion Sampling in Text-Aware Latent Space for Robust and Diverse Text-to-Music Generation

Taekoan Yoo, Wonkyung Jung, Kyunghun Kim, Kyeongbo Kong

ICML 2026 · Image generation › Autoregressive and latent generation · approach: Diffusion audio synthesis

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
perturb only padding positions (PADS) instead of full conditioning (CADS), with audio latent (ablation)SAO diffusion model · SongDescriber, matched CLAP-fVendi: 22.591 -> 24.367Improved
perturb only padding positions (PADS) instead of CADS, with audio latent (ablation)SAO diffusion model · SongDescriber, matched CLAP-fFDopenl3: 147.730 -> 139.838Improved
train diffusion in Text-Aware Latent space instead of audio latent, with CADS (ablation)SAO + CADS · MelBench, matched CLAP-fmRecall: 0.237 -> 0.391Improved
PADS in audio latent space without TAL (ablation)SAO + PADS · MelBench, matched CLAP-fmRecall: 0.237 -> 0.212Hurt
combine TAL and PADS versus audio latent with CADS (ablation)SAO · SongDescriber / MelBenchVendi: 22.591 -> 26.075 (15.4% higher diversity; 71.6% within-genre)Improved
add PADS on top of TAL (versus TAL with CADS) (ablation)TAL-DM · SongDescriber / MelBenchVendi: 23.873 -> 26.075Improved
apply PADS instead of CADS to ACE-StepACE-Step · SongDescriberCLAP-f: 0.1981 -> 0.2318Improved
apply PADS to MelodyFlowMelodyFlow · SongDescriberVendi: 18.2190 -> 21.5843Improved
add CADS perturbation to SAO (full conditioning noise)SAO · Human evaluation, alignment MOSMOS alignment: 3.83 -> 2.87Hurt
use PADS instead of CADS on SAOSAO · Human evaluation, within-modelMOS quality: 3.18 -> 3.89Improved
use PADS-TAL instead of SAO baselineSAO · Human evaluation, across-modelMOS diversity: 3.41 -> 3.71Improved
use step annealing schedule instead of linear in PADS (ablation)SAO + PADS · SongDescriberCLAP-f: 0.335 -> 0.312 CLAP-f; Vendi 23.230 -> 24.843Mixed

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