AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

BAR: Refactor the Basis of Autoregressive Visual Generation

Zhicong Tang, Dong Chen, Jianmin Bao, Baining Guo

ICLR 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images

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 learnable orthogonal basis transform (BAR) to xAR-HxAR-H · ImageNet 256x256FID: 1.24 -> 1.15Improved
add BAR to MAR-LMAR-L · ImageNet 256x256FID: 1.78 -> 1.56Improved
add BAR to MAR-HMAR-H · ImageNet 256x256FID: 1.55 -> 1.49Improved
add BAR to xAR-L at 512 resolutionxAR-L · ImageNet 512x512FID: 1.70 -> 1.63Improved
add BAR for text-to-imageBAR vs FAR · MS-COCO 2014 (FID-30K)FID: 13.91 -> 12.55Improved
initialize transform A with identity instead of random orthogonal (ablation)BAR on xAR-B · ImageNet 256x256FID: 1.66 -> 1.63Improved
apply soft orthogonal projection (delta 0.5) instead of none (ablation)BAR on xAR-B · ImageNet 256x256FID: 1.70 -> 1.63Improved
use soft instead of hard orthogonal projection (ablation)BAR on xAR-B · ImageNet 256x256FID: 1.66 -> 1.63Improved
add residual objective on top of L_BAR (ablation)BAR on xAR-B · ImageNet 256x256FID: 1.64 -> 1.63No clear effect
add learnable BAR to xAR-B with default settingsxAR-B · ImageNet 256x256FID: 1.72 -> 1.63Improved

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