AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Denoising without Diffusion: Fixed-Noise Denoiser Anomaly Detection in Tabular Data

Manuel Hirth, Lukas Koberg, Nasser Jazdi, Enkelejda Kasneci

ICML 2026 · Image generation › Fast diffusion sampling · approach: Denoising diffusion probabilistic models

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
use fixed noise without timestep embedding instead of EDM schedule (ablation)DenoiserAD · ADBenchAUCROC (%): 69.71 -> 83.34Improved
use fixed noise instead of noise schedule NCSN/NCSBAD (ablation)DenoiserAD · ADBenchAUCROC (%): 81.86 -> 83.34Improved
add timestep embedding to fixed noise (ablation)DenoiserAD · ADBenchAUCROC (%): 83.34 -> 83.11Hurt
sample noise from stable interval [0.12, 0.35] instead of fixed sigma^2 (ablation)DenoiserAD · ADBenchAUCROC (%): 83.34 -> 81.78Hurt
add explicit linear reference channel (vs plain DAE) with stability score S2 (ablation)DenoiserAD · ADBenchAUCROC (%): 82.81 -> 83.34Improved
use Monte-Carlo input-referenced score S2 instead of single-shot reconstruction S1 on plain DAE (ablation)DAE · ADBenchAUCROC (%): 82.38 -> 82.81 (+0.43)No clear effect
use stability score S2 instead of augmented reconstruction S4 (ablation)DenoiserAD · ADBenchAUCROC (%): 81.33 -> 83.34Improved
use stability score S2 instead of mean-centered variance S3 (ablation)DenoiserAD · ADBenchAUCROC (%): 75.88 -> 83.34Improved
use stability score S2 instead of Jacobian-norm score (ablation)DenoiserAD · ADBenchAUCROC (%): 81.05 -> 83.34Improved
apply fixed-noise stability formulation to DDPM-based AD (ablation)DDPM · ADBenchAUCROC (%): 70.73 -> 77.91Improved
apply fixed-noise stability formulation to DTE (ablation)DTE · ADBenchAUCROC (%): 79.77 -> 80.65Improved
replace DTE backbone with ResMLP only (ablation)DTE · ADBenchAUCROC (%): 79.77 -> 69.39Hurt

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