activity
20242026
collaborators

7 papers

cs.LG2026

Couple to Control: Joint Initial Noise Design in Diffusion Models

Jing Jia, Liyue Shen, Guanyang Wang

Diffusion models typically generate image batches from independent Gaussian initial noises. We argue that this independence assumption is only one choice within a broader class of…

cs.LG2026

Antithetic Noise in Diffusion Models

Jing Jia, Sifan Liu, Bowen Song +3

We systematically study antithetic initial noise in diffusion models, discovering that pairing each noise sample with its negation consistently produces strong negative correlation…

cs.CV2025

Local Patches Meet Global Context: Scalable 3D Diffusion Priors for Computed Tomography Reconstruction

Taewon Yang, Jason Hu, Jeffrey A. Fessler +1

Diffusion models learn strong image priors that can be leveraged to solve inverse problems like medical image reconstruction. However, for real-world applications such as 3D Comput…

cs.CV2025

Learning Image Priors through Patch-based Diffusion Models for Solving Inverse Problems

Jason Hu, Bowen Song, Xiaojian Xu +2

Diffusion models can learn strong image priors from underlying data distribution and use them to solve inverse problems, but the training process is computationally expensive and r…

cs.LG2025

CCS: Controllable and Constrained Sampling with Diffusion Models via Initial Noise Perturbation

Bowen Song, Zecheng Zhang, Zhaoxu Luo +6

Diffusion models have emerged as powerful tools for generative tasks, producing high-quality outputs across diverse domains. However, how the generated data responds to the initial…

cs.CV2024

SatDiffMoE: A Mixture of Estimation Method for Satellite Image Super-resolution with Latent Diffusion Models

Zhaoxu Luo, Bowen Song, Liyue Shen

During the acquisition of satellite images, there is generally a trade-off between spatial resolution and temporal resolution (acquisition frequency) due to the onboard sensors of…