Correcting Diffusion Generation through Resampling

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Yujian, Zhang, Yang, Jaakkola, Tommi, Chang, Shiyu
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909337908150272
author Liu, Yujian
Zhang, Yang
Jaakkola, Tommi
Chang, Shiyu
author_facet Liu, Yujian
Zhang, Yang
Jaakkola, Tommi
Chang, Shiyu
contents Despite diffusion models' superior capabilities in modeling complex distributions, there are still non-trivial distributional discrepancies between generated and ground-truth images, which has resulted in several notable problems in image generation, including missing object errors in text-to-image generation and low image quality. Existing methods that attempt to address these problems mostly do not tend to address the fundamental cause behind these problems, which is the distributional discrepancies, and hence achieve sub-optimal results. In this paper, we propose a particle filtering framework that can effectively address both problems by explicitly reducing the distributional discrepancies. Specifically, our method relies on a set of external guidance, including a small set of real images and a pre-trained object detector, to gauge the distribution gap, and then design the resampling weight accordingly to correct the gap. Experiments show that our methods can effectively correct missing object errors and improve image quality in various image generation tasks. Notably, our method outperforms the existing strongest baseline by 5% in object occurrence and 1.0 in FID on MS-COCO. Our code is publicly available at https://github.com/UCSB-NLP-Chang/diffusion_resampling.git.
format Preprint
id arxiv_https___arxiv_org_abs_2312_06038
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Correcting Diffusion Generation through Resampling
Liu, Yujian
Zhang, Yang
Jaakkola, Tommi
Chang, Shiyu
Computer Vision and Pattern Recognition
Machine Learning
Despite diffusion models' superior capabilities in modeling complex distributions, there are still non-trivial distributional discrepancies between generated and ground-truth images, which has resulted in several notable problems in image generation, including missing object errors in text-to-image generation and low image quality. Existing methods that attempt to address these problems mostly do not tend to address the fundamental cause behind these problems, which is the distributional discrepancies, and hence achieve sub-optimal results. In this paper, we propose a particle filtering framework that can effectively address both problems by explicitly reducing the distributional discrepancies. Specifically, our method relies on a set of external guidance, including a small set of real images and a pre-trained object detector, to gauge the distribution gap, and then design the resampling weight accordingly to correct the gap. Experiments show that our methods can effectively correct missing object errors and improve image quality in various image generation tasks. Notably, our method outperforms the existing strongest baseline by 5% in object occurrence and 1.0 in FID on MS-COCO. Our code is publicly available at https://github.com/UCSB-NLP-Chang/diffusion_resampling.git.
title Correcting Diffusion Generation through Resampling
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2312.06038