AI Paper Atlas

Research map / Image generation / Image restoration

Query-Based Asymmetric Modeling with Decoupled Input–Output Rates for Speech Restoration

Ui-Hyeop Shin, Jaehyun Ko, Woocheol Jeong, Hyung-Min Park

ICML 2026 · Image generation › Image restoration · approach: Efficient restoration architectures

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 decoder with cross-attention (MHCA) to encoder-only model (ablation)TF-Restormer · VCTK SSR 8->16 kHzLSD: 2.12 -> 0.89Improved
add MHCA cross-attention in decoder instead of self-attention only (ablation)TF-Restormer · VCTK SSR 8->44.1 kHzLSD: 1.35 -> 1.01Improved
use size-matched small encoder-decoder (10.9M) instead of encoder-only (11.6M) (ablation)TF-Restormer · VCTK SSR 8->16 kHzLSD: 2.12 -> 1.36Improved
use encoder-decoder instead of encoder-only (ablation)TF-Restormer · VCTK SSR 8->44.1 kHzMACs (G): 415.1 -> 308.4Improved
use shared SFI-STFT discriminator instead of separate rate-specific discriminators (ablation)TF-Restormer · VCTK+DEMAND DNPESQ: 3.28 -> 3.41Improved
add shared frequency projection layer (ablation)TF-Restormer · UNIVERSE GSRUTMOS: 3.90 -> 4.10Improved
use separate (per-module) frequency projections instead of shared (ablation)TF-Restormer · VCTK SSR 8->16 kHzPESQ: 3.70 -> 3.55Hurt
add spectral loss on top of perceptual loss alone (ablation)TF-Restormer pretraining · UNIVERSE GSRPESQ: 1.85 -> 2.23Improved
replace l1 with l2 on complex STFT components (ablation)TF-Restormer pretraining · UNIVERSE GSRPESQ: 2.23 -> 2.21Hurt
use plain log1p (w=1) instead of l1 (ablation)TF-Restormer pretraining · UNIVERSE GSRPESQ: 2.23 -> 2.25No clear effect
use adaptive w_tf scaled log-spectral loss instead of fixed scale w=1e-3 (ablation)TF-Restormer pretraining · VCTK SSR 8->16 kHzMCD: 1.37 -> 1.29Improved
use overly small scale w=1e-4 in s-log1p (ablation)TF-Restormer pretraining · VCTK SSR 8->16 kHzMCD: 1.37 -> 2.43Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 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