AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Learning on the Manifold: Unlocking Standard Diffusion Transformers with Representation Encoders

Amandeep Kumar, Vishal Patel

ECCV 2026 · Image generation › Autoregressive and latent generation · approach: High-resolution diffusion synthesis · 15 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 Euclidean flow matching with Riemannian flow matching (SLERP geodesic) (ablation)DiT-B/1 on DINOv2-B · ImageNet 256FID: 24.32 -> 7.06Improved
project noise and latents to the unit sphere only (ablation)DiT-B/1 on DINOv2-B · ImageNet 256FID: 24.32 -> 21.99No clear effect
add Jacobi regularization on top of RFM (ablation)DiT-B/1 on DINOv2-B · ImageNet 256FID: 7.06 -> 6.77Improved
train RJF longer (80 to 200 epochs) (ablation)DiT-B/1 RJF · ImageNet 256FID: 6.77 -> 4.95Improved
add RJF to LightningDiT-XL (DINOv2-B)LightningDiT-XL · ImageNet 256FID: 4.28 -> 3.62Improved
add RJF to DDT-XL/1DDT-XL · ImageNet 256FID: 6.55 -> 5.82Improved
add RJF to width-scaled DiT^DHDiT^DH · ImageNet 256FID: 6.33 -> 6.20Improved
add RJF with SigLIP latentsDiT-B/1 · ImageNet 256 (SigLIP)FID: 130.21 -> 10.39Improved
add RJF with MAE latentsDiT-B/1 · ImageNet 256 (MAE)FID: 50.48 -> 19.82Improved
increase inference projection radius from original norm to R=45 (ablation)DiT-B/1 RJF · ImageNet 256FID: 7.79 -> 6.77Improved
compare RJF with guidance to SFD at 80 epochsDiT-XL RJF · ImageNet 256 (guidance)FID: 2.81 vs 1.30 (SFD)Hurt
mask the radial loss and train only angular component (width sweep) (ablation)DiT-S single image overfit · DINOv2 tokens d=768angular loss convergenceImproved

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