Spanning Tree Autoregressive Visual Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Sangkyu, Lee, Changho, Han, Janghoon, Song, Hosung, You, Tackgeun, Lim, Hwasup, Choi, Stanley Jungkyu, Lee, Honglak, Yu, Youngjae
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918213190680576
author Lee, Sangkyu
Lee, Changho
Han, Janghoon
Song, Hosung
You, Tackgeun
Lim, Hwasup
Choi, Stanley Jungkyu
Lee, Honglak
Yu, Youngjae
author_facet Lee, Sangkyu
Lee, Changho
Han, Janghoon
Song, Hosung
You, Tackgeun
Lim, Hwasup
Choi, Stanley Jungkyu
Lee, Honglak
Yu, Youngjae
contents We present Spanning Tree Autoregressive (STAR) modeling, which can incorporate prior knowledge of images, such as center bias and locality, to maintain sampling performance while also providing sufficiently flexible sequence orders to accommodate image editing at inference. Approaches that expose randomly permuted sequence orders to conventional autoregressive (AR) models in visual generation for bidirectional context either suffer from a decline in performance or compromise the flexibility in sequence order choice at inference. Instead, STAR utilizes traversal orders of uniform spanning trees sampled in a lattice defined by the positions of image patches. Traversal orders are obtained through breadth-first search, allowing us to efficiently construct a spanning tree whose traversal order ensures that the connected partial observation of the image appears as a prefix in the sequence through rejection sampling. Through the tailored yet structured randomized strategy compared to random permutation, STAR preserves the capability of postfix completion while maintaining sampling performance without any significant changes to the model architecture widely adopted in the language AR modeling.
format Preprint
id arxiv_https___arxiv_org_abs_2511_17089
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Spanning Tree Autoregressive Visual Generation
Lee, Sangkyu
Lee, Changho
Han, Janghoon
Song, Hosung
You, Tackgeun
Lim, Hwasup
Choi, Stanley Jungkyu
Lee, Honglak
Yu, Youngjae
Computer Vision and Pattern Recognition
Artificial Intelligence
We present Spanning Tree Autoregressive (STAR) modeling, which can incorporate prior knowledge of images, such as center bias and locality, to maintain sampling performance while also providing sufficiently flexible sequence orders to accommodate image editing at inference. Approaches that expose randomly permuted sequence orders to conventional autoregressive (AR) models in visual generation for bidirectional context either suffer from a decline in performance or compromise the flexibility in sequence order choice at inference. Instead, STAR utilizes traversal orders of uniform spanning trees sampled in a lattice defined by the positions of image patches. Traversal orders are obtained through breadth-first search, allowing us to efficiently construct a spanning tree whose traversal order ensures that the connected partial observation of the image appears as a prefix in the sequence through rejection sampling. Through the tailored yet structured randomized strategy compared to random permutation, STAR preserves the capability of postfix completion while maintaining sampling performance without any significant changes to the model architecture widely adopted in the language AR modeling.
title Spanning Tree Autoregressive Visual Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2511.17089