Improved Constrained Generation by Bridging Pretrained Generative Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liang, Xiaoxuan, Naderiparizi, Saeid, Liu, Yunpeng, Zwartsenberg, Berend, Wood, Frank
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917319426441216
author Liang, Xiaoxuan
Naderiparizi, Saeid
Liu, Yunpeng
Zwartsenberg, Berend
Wood, Frank
author_facet Liang, Xiaoxuan
Naderiparizi, Saeid
Liu, Yunpeng
Zwartsenberg, Berend
Wood, Frank
contents Constrained generative modeling is fundamental to applications such as robotic control and autonomous driving, where models must respect physical laws and safety-critical constraints. In real-world settings, these constraints rarely take the form of simple linear inequalities, but instead complex feasible regions that resemble road maps or other structured spatial domains. We propose a constrained generation framework that generates samples directly within such feasible regions while preserving realism. Our method fine-tunes a pretrained generative model to enforce constraints while maintaining generative fidelity. Experimentally, our method exhibits characteristics distinct from existing fine-tuning and training-free constrained baselines, revealing a new compromise between constraint satisfaction and sampling quality.
format Preprint
id arxiv_https___arxiv_org_abs_2603_06742
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Improved Constrained Generation by Bridging Pretrained Generative Models
Liang, Xiaoxuan
Naderiparizi, Saeid
Liu, Yunpeng
Zwartsenberg, Berend
Wood, Frank
Machine Learning
Artificial Intelligence
Robotics
Constrained generative modeling is fundamental to applications such as robotic control and autonomous driving, where models must respect physical laws and safety-critical constraints. In real-world settings, these constraints rarely take the form of simple linear inequalities, but instead complex feasible regions that resemble road maps or other structured spatial domains. We propose a constrained generation framework that generates samples directly within such feasible regions while preserving realism. Our method fine-tunes a pretrained generative model to enforce constraints while maintaining generative fidelity. Experimentally, our method exhibits characteristics distinct from existing fine-tuning and training-free constrained baselines, revealing a new compromise between constraint satisfaction and sampling quality.
title Improved Constrained Generation by Bridging Pretrained Generative Models
topic Machine Learning
Artificial Intelligence
Robotics
url https://arxiv.org/abs/2603.06742