Diffusion probabilistic models (DPMs) represent a class of powerful generative models. Despite their success, the inference of DPMs is expensive since it generally needs to iterate over thousands of timesteps. A key problem in the inference is to estimate the variance in each timestep of the reverse process. In this work, we present a surprising result that both the optimal reverse variance and the corresponding optimal KL divergence of a DPM have analytic forms w.r.t. its score function. Building upon it, we propose Analytic-DPM, a training-free inference framework that estimates the analytic forms of the variance and KL divergence using the Monte Carlo method and a pretrained score-based model. Further, to correct the potential bias caused by the score-based model, we derive both lower and upper bounds of the optimal variance and clip the estimate for a better result. Empirically, our analytic-DPM improves the log-likelihood of various DPMs, produces high-quality samples, and meanwhile enjoys a 20x to 80x speed up.
Department of Computer Science, Tsinghua University(清华大学计算机科学与技术系);Sea AI Lab, Singapore(Sea AI实验室,新加坡);Qing Yuan Research Institute, Shanghai Jiao Tong University(上海交通大学清源研究院)
Disentangling Total-Variance and Signal-to-Noise-Ratio Improves Diffusion Models
BIFOLD – Berlin Institute for the Foundations of Learning and Data Machine Learning Group, Technische Universität Berlin(柏林学习与数据基础研究所机器学习组,柏林技术大学);Technische Universität Berlin(柏林技术大学);Google DeepMind(谷歌DeepMind);Department of Artificial Intelligence, Korea University(韩国大学人工智能系);Max-Planck Institute for Informatics(马克斯·普朗克信息研究所);RIKEN Center for Advanced Intelligence Project(理化学研究所高级智能项目中心)
Kernel-Smoothed Scores for Denoising Diffusion: A Bias-Variance Study
Université Claude Bernard Lyon 1(克莱尔-伯纳德里昂1大学);Laboratoire SAF, ISFA, France(SAF实验室,ISFA,法国);Department of Mathematics University of Vienna, Austria(数学系,维也纳大学,奥地利);The Ohio State University, USA(俄亥俄州立大学,美国)
SA-Solver: Stochastic Adams Solver for Fast Sampling of Diffusion Models
University of Chinese Academy of Sciences(中国科学院大学);Huawei Noah’s Ark Lab(华为诺亚实验室);Peking University(北京大学);Academy of Mathematics and Systems Science, Chinese Academy of Sciences(中国科学院数学与系统科学研究院)
Generative Diffusion Surrogates with Analytical Variance Schedule
University of Oxford(牛津大学);Princeton University(普林斯顿大学);Max Planck Institute for Security and Privacy(马克斯·普朗克安全与隐私研究所);University of Cambridge(剑桥大学);Gonville & Caius College(贡维尔与凯斯学院)
Mitigating Privacy Risk in Membership Inference by Convex-Concave Loss
Southern University of Science and Technology(南方科技大学);Singapore University of Technology and Design(新加坡科技设计大学);South China University of Technology(华南理工大学);Jilin University(吉林大学)
Designing Reinforcement Learning for Diffusion Models: A Unified Path-Space View
State Key Laboratory of General Artificial Intelligence(通用人工智能国家重点实验室);School of Intelligence Science and Technology, Peking University(北京大学智能科学与技术学院);ByteDance(字节跳动)