AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

VA-p: Variational Policy Alignment for Pixel-Aware Autoregressive Generation

Xinyao Liao, Qiyuan He, Kai Xu, Xiaoye Qu, Yicong Li, Wei Wei, Angela Yao

CVPR 2026 · Image generation › Text-to-image diffusion · approach: T2I evaluation and controllable adapters

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 VA-pi post-training (pixel reconstruction reward plus prior regularization)LlamaGen-XXL · ImageNet-1K w/o CFGFID: 14.36 -> 7.65Improved
add VA-pi post-trainingLlamaGen-XL · GenEvalOverall: 0.306 -> 0.339Improved
add VA-pi post-trainingJanus-Pro 1B · GenEvalOverall: 0.725 -> 0.744Improved
use VA-pi instead of external-reward AR-GRPOLlamaGen-XL · GenEvalOverall: 0.324 -> 0.339Improved
add VA-pi post-training on LlamaGen-XL with classifier-free guidanceLlamaGen-XL · ImageNet-1K w/ CFGFID: 2.79 -> 2.94Hurt
add reconstruction reward without prior regularization (ablation)VA-pi · ImageNet-1K w/o CFGFID: 14.36 -> 38.63Hurt
add prior regularization (cross-entropy) alone without reconstruction reward (ablation)VA-pi · ImageNet-1K w/o CFGFID: 14.36 -> 14.17No clear effect
combine reconstruction reward with prior regularization (ablation)VA-pi · ImageNet-1K w/o CFGFID: 38.63 -> 7.65Improved
use contextual noise xi=0.5 instead of 0 in LlamaGen T2I post-training (ablation)VA-pi · GenEvalOverall: 0.302 -> 0.339Improved
increase contextual noise above 0.75 (ablation)VA-pi · GenEvalOverall: 0.339 -> 0.329 (xi=0.95)Hurt
use cross-entropy instead of KL regularization with beta 0.1 (ablation)VA-pi · ImageNet-1K w/o CFGFID and ISImproved
remove regularization (beta=0) or set it too strong (beta=1.0) (ablation)VA-pi · ImageNet-1K w/o CFGFID: FID 38.63 without regularizationHurt

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