arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2403.01666cs.LGcs.CV
收藏

通过扩散过程改进对抗能量模型

Improving Adversarial Energy-Based Model via Diffusion Process

vivo Mobile Communication Co., Ltd(维沃移动通信有限公司);Stevens Institute of Technology(史蒂文斯理工学院);Technical University of Denmark(丹麦技术大学)

Cong Geng, Tian Han, Peng-Tao Jiang, Hao Zhang, Jinwei Chen, Søren Hauberg, Bo Li

AI总结:

本文将能量模型嵌入扩散去噪步骤,并结合对称Jeffrey散度与变分后验分布改进对抗能量模型的训练,在提升生成质量的同时支持高效密度估计。

摘要:

生成模型已展现出强大的生成能力,但对高效似然估计的探索较少。能量模型(EBMs)定义灵活的能量函数以高效地参数化未归一化密度,但以难以训练而著称。对抗能量模型引入生成器形成极小极大训练博弈,以避免传统能量模型所使用的昂贵MCMC采样,但对抗能量模型与其他强生成模型之间仍存在明显差距。受基于扩散的模型启发,我们将能量模型嵌入每个去噪步骤中,将漫长的生成过程拆分为若干较小步骤。此外,我们采用对称Jeffrey散度,并为生成器的训练引入变分后验分布,以解决对抗能量模型中存在的主要挑战。我们的实验表明,与现有对抗能量模型相比,生成质量有显著提升,同时还提供了有用的能量函数以实现高效密度估计。

英文摘要:

Generative models have shown strong generation ability while efficient likelihood estimation is less explored. Energy-based models~(EBMs) define a flexible energy function to parameterize unnormalized densities efficiently but are notorious for being difficult to train. Adversarial EBMs introduce a generator to form a minimax training game to avoid expensive MCMC sampling used in traditional EBMs, but a noticeable gap between adversarial EBMs and other strong generative models still exists. Inspired by diffusion-based models, we embedded EBMs into each denoising step to split a long-generated process into several smaller steps. Besides, we employ a symmetric Jeffrey divergence and introduce a variational posterior distribution for the generator's training to address the main challenges that exist in adversarial EBMs. Our experiments show significant improvement in generation compared to existing adversarial EBMs, while also providing a useful energy function for efficient density estimation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑