Chang Wan

Chang Wan

Auto-generated profile, based on publicly available information

Chang Wan

Zhejiang Normal University, Jinhua, Zhejiang, China

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
9 May 2016 in Advanced Materials
Verified
Journal Article
Closed access
PDF file
CWChang Jin Wan
Chang Jin Wan
YLYanghui Liu
Yanghui Liu
PFPing Feng
Ping Feng
WWWei Wang
Wei Wang
LZLiqiang Zhu
Liqiang Zhu
QWQing Wan
Qing Wan

11 September 2024 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
CWChang Wan
Chang Wan
MYMing Yang
Ming Yang
MLMinglu Li
Minglu Li
YJYunliang Jiang
Yunliang Jiang
ZZZl Zheng
Zl Zheng
Recently, researchers have proposed many deep generative models, including generative adversarial networks (GANs) and denoising diffusion models. Although significant breakthroughs have been made and empirical success has been achieved with the GAN, its mathematical underpinnings remain relatively unknown. This paper focuses on a rigorous mathematical theoretical framework: the composite-functional-gradient GAN (CFG) [1]. Specifically, we reveal the theoretical connection between the CFG model and score-based models. We find that the CFG discriminator’s training objective is equivalent to finding an optimal D (x). The optimal D (x)’s gradient differentiates the integral of the differences between the score functions of real and synthesized samples. Conversely, training the CFG generator involves finding an optimal G (x) that minimizes this difference. In this paper, we aim to derive an annealed weight preceding the CFG discriminator’s weight. This new explicit theoretical explanation model is called the annealed CFG method. To overcome the annealed CFG method’s limitation, as the method is not readily applicable to the state-of-the-art (SOTA) GAN model, we propose a nested annealed training scheme (NATS). This scheme keeps the annealed weight from the CFG method and can be seamlessly adapted to various GAN models, no matter their structural, loss, or regularization differences. We conduct thorough experimental evaluations on various benchmark datasets for image generation. The results show that our annealed CFG and NATS methods significantly improve the synthesized samples’ quality and diversity. This improvement is clear when comparing the CFG method and the SOTA GAN models.

10 January 2025
Verified
Conference Paper
Open access
XHXiaoyu Hu
Xiaoyu Hu
CWChang Wan
Chang Wan
ZWZhigang Wang
Zhigang Wang