AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Drift-AR: Single-Step Visual Autoregressive Generation via Anti-Symmetric Drifting

zhen zou, Xiaoxiao Ma, Mingde Yao, Jie Huang, Linjiang Huang, Feng Zhao

ECCV 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images · 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
use entropy-parameterized variance instead of constant variance (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)FID: 1.72 -> 1.57Improved
use entropy-parameterized variance instead of constant variance (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)IS: 289.3 -> 298.1Improved
use annealed alpha(t) instead of fixed alpha=0.5 (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)FID: 1.69 -> 1.57Improved
add repulsion term (anti-symmetric kernel objective) (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)FID: 1.69 -> 1.57Improved
add entropy-based early stopping in speculative decoding (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)FID: 1.62 -> 1.57Improved
freeze AR prior in Phase II (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)FID: 1.67 -> 1.57Improved
use drifting decoder instead of diffusion decoder at 1 step (ablation)TransDiff-H · ImageNet 256x256 (TransDiff-H) 1 stepFID: 14.72 -> 1.57Improved
use drifting decoder instead of DMD at 1 step (ablation)TransDiff-H · ImageNet 256x256 (TransDiff-H) 1 stepFID: 2.93 -> 1.57Improved
sigma_max 1.0 instead of 0.5 (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)FID: 1.57 -> 1.64Hurt
sigma_max 0.1 instead of 0.5 (ablation)Drift-AR · ImageNet 256x256 (TransDiff-H)FID: 1.57 -> 1.68Hurt
apply Drift-AR to NextStep-1NextStep-1 · GenEval / MJHQ-30KFID: 6.71 -> 6.66Improved
apply DMD to NextStep-1NextStep-1 · GenEval / MJHQ-30KFID: 6.71 -> 8.33Hurt

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