M-EBM: Towards Understanding the Manifolds of Energy-Based Models

Advances in Knowledge Discovery and Data Mining : 21st Pacific-Asia Conference, PAKDD 2017, Jeju, South Korea, May 23-26, 2017, Proceedings. Part I. Pacific-Asia Conference on Knowledge Discovery and Data Mining (21st : 2017 : Cheju Isl... Pub Date : 2023-03-08 DOI:10.48550/arXiv.2303.04343

Xiulong Yang, Shihao Ji

{"title":"M-EBM: Towards Understanding the Manifolds of Energy-Based Models","authors":"Xiulong Yang, Shihao Ji","doi":"10.48550/arXiv.2303.04343","DOIUrl":null,"url":null,"abstract":"Energy-based models (EBMs) exhibit a variety of desirable properties in predictive tasks, such as generality, simplicity and compositionality. However, training EBMs on high-dimensional datasets remains unstable and expensive. In this paper, we present a Manifold EBM (M-EBM) to boost the overall performance of unconditional EBM and Joint Energy-based Model (JEM). Despite its simplicity, M-EBM significantly improves unconditional EBMs in training stability and speed on a host of benchmark datasets, such as CIFAR10, CIFAR100, CelebA-HQ, and ImageNet 32x32. Once class labels are available, label-incorporated M-EBM (M-JEM) further surpasses M-EBM in image generation quality with an over 40% FID improvement, while enjoying improved accuracy. The code can be found at https://github.com/sndnyang/mebm.","PeriodicalId":91995,"journal":{"name":"Advances in Knowledge Discovery and Data Mining : 21st Pacific-Asia Conference, PAKDD 2017, Jeju, South Korea, May 23-26, 2017, Proceedings. Part I. Pacific-Asia Conference on Knowledge Discovery and Data Mining (21st : 2017 : Cheju Isl...","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2023-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Advances in Knowledge Discovery and Data Mining : 21st Pacific-Asia Conference, PAKDD 2017, Jeju, South Korea, May 23-26, 2017, Proceedings. Part I. Pacific-Asia Conference on Knowledge Discovery and Data Mining (21st : 2017 : Cheju Isl...","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.48550/arXiv.2303.04343","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 0

Abstract

Energy-based models (EBMs) exhibit a variety of desirable properties in predictive tasks, such as generality, simplicity and compositionality. However, training EBMs on high-dimensional datasets remains unstable and expensive. In this paper, we present a Manifold EBM (M-EBM) to boost the overall performance of unconditional EBM and Joint Energy-based Model (JEM). Despite its simplicity, M-EBM significantly improves unconditional EBMs in training stability and speed on a host of benchmark datasets, such as CIFAR10, CIFAR100, CelebA-HQ, and ImageNet 32x32. Once class labels are available, label-incorporated M-EBM (M-JEM) further surpasses M-EBM in image generation quality with an over 40% FID improvement, while enjoying improved accuracy. The code can be found at https://github.com/sndnyang/mebm.

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

M-EBM:迈向理解基于能源模型的流形

基于能量的模型(EBMs)在预测任务中表现出各种理想的特性，例如通用性、简单性和组合性。然而，在高维数据集上训练EBMs仍然不稳定且昂贵。在本文中，我们提出了一种流形实证模型(M-EBM)来提高无条件实证模型和联合能量模型(JEM)的整体性能。尽管M-EBM很简单，但它在一系列基准数据集(如CIFAR10、CIFAR100、CelebA-HQ和ImageNet 32x32)上显著提高了无条件ebm的训练稳定性和速度。一旦类标签可用，包含标签的M-EBM (M-JEM)在图像生成质量上进一步超过M-EBM, FID改进超过40%，同时精度也有所提高。代码可以在https://github.com/sndnyang/mebm上找到。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

Advances in Knowledge Discovery and Data Mining : 21st Pacific-Asia Conference, PAKDD 2017, Jeju, South Korea, May 23-26, 2017, Proceedings. Part I. Pacific-Asia Conference on Knowledge Discovery and Data Mining (21st : 2017 : Cheju Isl...

自引率

0.00%

发文量