AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Optimality of FSQ Tokens for Continuous Diffusion for Categorical Data with Application to Text-to-Speech

Vadim Popov, Wenju Gu, Tasnima Sadekova, Georgii Aparin, Assel Yermekova

ICML 2026 · Image generation › Autoregressive and latent generation · approach: Diffusion audio synthesis · 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
replace CosyVoice2's 0.5B LLM text-to-token module with a 45M CDCD DiT on FSQ tokensCosyVoice2 · SEED-TTS test-enWER: 2.57% -> 2.00%Improved
use true FSQ embeddings instead of perturbed embeddings (ablation)CDCD-TTS · SEED-TTS test-en, 5 stepsWER: 3.10% -> 2.39%Improved
use true FSQ embeddings instead of perturbed embeddings with many denoising steps (ablation)CDCD-TTS · SEED-TTS test-en, 25 stepsWER: 2.03% -> 2.00%No clear effect
keep FSQ embeddings but permute token ids (ablation)CDCD-TTS · SEED-TTS test-en, 25 stepsWER: 2.00% -> 15.4%Hurt
use RVQ 128-dim EnCodec tokens instead of FSQ 8-dim tokens (ablation)CDCD-TTS · SEED-TTS test-en, 25 stepsWER: 2.00% -> 21.3%Hurt
use true FSQ latent space instead of perturbed in toy unconditional generation (base2) (ablation)CDCD · Toy sequences, base2log-KL: -8.68 -> -9.55Improved
use true FSQ latent space instead of perturbed in toy unconditional generation (base3) (ablation)CDCD · Toy sequences, base3log-KL: -6.72 -> -7.16Improved
use FSQ codebook instead of random codebooks for optimal-model token accuracy (base3, 2D) (ablation)CDCD latent space · Monte Carlo, 1000 random codebooksprediction accuracy: 16.659% -> 17.052%Improved
increase denoising steps from 5 to 25 for FSQ-original (ablation)CDCD-TTS · SEED-TTS test-enWER: 2.39% -> 2.00%Improved
CDCD-TTS vs F5-TTS reference model on WERCDCD-TTS · SEED-TTS test-enWER: 1.83% (F5-TTS) vs 2.00%Hurt

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