中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
MoEP-AE: Autoencoding Mixtures of Exponential Power Distributions for Open-Set Recognition

文献类型：期刊论文

作者Sun, Jiayin2,3,4; Wang, Hong1; Dong, Qiulei2,3,4; Qiulei, Dong; Jiayin, Sun
刊名IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
出版日期2023
卷号33期号:1页码:312-325
关键词Feature extraction Task analysis Training Power distribution Sun Decoding Gaussian distribution Open-set recognition autoencoder scale mixture distribution exponential power distribution
ISSN号1051-8215
DOI10.1109/TCSVT.2022.3200112
通讯作者Dong, Qiulei(qldong@nlpr.ia.ac.cn)
英文摘要Open-set recognition aims to identify unknown classes while maintaining classification performance on known classes and has attracted increasing attention in the pattern recognition field. However, how to learn effective feature representations whose distributions are usually complex for classifying both known-class and unknown-class samples when only the known-class samples are available for training is an ongoing issue in open-set recognition. In contrast to methods implementing a single Gaussian, a mixture of Gaussians (MoG), or multiple MoGs, we propose a novel autoencoder that learns feature representations by modeling them as mixtures of exponential power distributions (MoEPs) in latent spaces called MoEP-AE. The proposed autoencoder considers that many real-world distributions are sub-Gaussian or super-Gaussian and can thus be represented by MoEPs rather than a single Gaussian or an MoG or multiple MoGs. We design a differentiable sampler that can sample from an MoEP to guarantee that the proposed autoencoder is trained effectively. Furthermore, we propose an MoEP-AE-based method for open-set recognition by introducing a discrimination strategy, where the MoEP-AE is used to model the distributions of the features extracted from the input known-class samples by minimizing a designed loss function at the training stage, called MoEP-AE-OSR. Extensive experimental results in both standard-dataset and cross-dataset settings demonstrate that the MoEP-AE-OSR method outperforms 14 existing open-set recognition methods in most cases in both open-set recognition and closed-set recognition tasks.
资助项目National Key Research and Development Program of China[2021ZD0201600] ; National Natural Science Foundation of China[U1805264] ; National Natural Science Foundation of China[61972374] ; National Natural Science Foundation of China[61991423] ; Strategic Priority Research Program of the Chinese Academy of Sciences[XDB32050100] ; Beijing Municipal Science and Technology Project[Z211100011021004]
WOS研究方向Engineering
语种英语
WOS记录号WOS:000911746000023
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
资助机构National Key Research and Development Program of China ; National Natural Science Foundation of China ; Strategic Priority Research Program of the Chinese Academy of Sciences ; Beijing Municipal Science and Technology Project
源URL[http://ir.ia.ac.cn/handle/173211/51356]  
专题多模态人工智能系统全国重点实验室
通讯作者Dong, Qiulei; Qiulei, Dong
作者单位1.Univ Chinese Acad Sci, Coll Life Sci, Beijing 100049, Peoples R China
2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China
3.Chinese Acad Sci, Ctr Excellence Brain Sci & Intelligence Technol, Beijing 100190, Peoples R China
4.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
推荐引用方式
GB/T 7714		 Sun, Jiayin,Wang, Hong,Dong, Qiulei,et al. MoEP-AE: Autoencoding Mixtures of Exponential Power Distributions for Open-Set Recognition[J]. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY,2023,33(1):312-325.
APA Sun, Jiayin,Wang, Hong,Dong, Qiulei,Qiulei, Dong,&Jiayin, Sun.(2023).MoEP-AE: Autoencoding Mixtures of Exponential Power Distributions for Open-Set Recognition.IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY,33(1),312-325.
MLA Sun, Jiayin,et al."MoEP-AE: Autoencoding Mixtures of Exponential Power Distributions for Open-Set Recognition".IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 33.1(2023):312-325.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。