Out-Of-Distribution Detection In Unsupervised Continual Learning

被引:2
|
作者
He, Jiangpeng [1 ]
Zhu, Fengqing [1 ]
机构
[1] Purdue Univ, Elmore Family Sch Elect & Comp Engn, W Lafayette, IN 47907 USA
关键词
D O I
10.1109/CVPRW56347.2022.00430
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Unsupervised continual learning aims to learn new tasks incrementally without requiring human annotations. However, most existing methods, especially those targeted on image classification, only work in a simplified scenario by assuming all new data belong to new tasks, which is not realistic if the class labels are not provided. Therefore, to perform unsupervised continual learning in real life applications, an out-of-distribution detector is required at beginning to identify whether each new data corresponds to a new task or already learned tasks, which still remains under-explored yet. In this work, we formulate the problem for Out-of-distribution Detection in Unsupervised Continual Learning (OOD-UCL) with the corresponding evaluation protocol. In addition, we propose a novel OOD detection method by correcting the output bias at first and then enhancing the output confidence for in-distribution data based on task discriminativeness, which can be applied directly without modifying the learning procedures and objectives of continual learning. Our method is evaluated on CIFAR-100 dataset by following the proposed evaluation protocol and we show improved performance compared with existing OOD detection methods under the unsupervised continual learning scenario.
引用
收藏
页码:3849 / 3854
页数:6
相关论文
共 50 条
  • [41] Combining Contrastive Learning with Auto-Encoder for Out-of-Distribution Detection
    Luo, Dawei
    Zhou, Heng
    Bae, Joonsoo
    Yun, Bom
    APPLIED SCIENCES-BASEL, 2023, 13 (23):
  • [42] Learning by Erasing: Conditional Entropy Based Transferable Out-of-Distribution Detection
    Xing, Meng
    Feng, Zhiyong
    Su, Yong
    Oh, Changjae
    THIRTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 38 NO 6, 2024, : 6261 - 6269
  • [43] In or Out? Fixing ImageNet Out-of-Distribution Detection Evaluation
    Bitterwolf, Julian
    Mueller, Maximilian
    Hein, Matthias
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 202, 2023, 202
  • [44] DEEP DIRICHLET UNCERTAINTY FOR UNSUPERVISED OUT-OF-DISTRIBUTION DETECTION OF EYE FUNDUS PHOTOGRAPHS IN GLAUCOMA SCREENING
    Araojo, Teresa
    Aresta, Guilherme
    Bogunovic, Hrvoje
    2022 IEEE INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING CHALLENGES (IEEE ISBI 2022), 2022,
  • [45] Unsupervised Out-of-Distribution Object Detection via PCA-Driven Dynamic Prototype Enhancement
    Wu, Aming
    Deng, Cheng
    Liu, Wei
    IEEE TRANSACTIONS ON IMAGE PROCESSING, 2024, 33 : 2431 - 2446
  • [46] Towards In-Distribution Compatible Out-of-Distribution Detection
    Wu, Boxi
    Jiang, Jie
    Ren, Haidong
    Du, Zifan
    Wang, Wenxiao
    Li, Zhifeng
    Cai, Deng
    He, Xiaofei
    Lin, Binbin
    Liu, Wei
    THIRTY-SEVENTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 37 NO 9, 2023, : 10333 - 10341
  • [47] Out-of-Distribution Detection Using Outlier Detection Methods
    Diers, Jan
    Pigorsch, Christian
    IMAGE ANALYSIS AND PROCESSING, ICIAP 2022, PT III, 2022, 13233 : 15 - 26
  • [48] On the Impact of Spurious Correlation for Out-of-Distribution Detection
    Ming, Yifei
    Yin, Hang
    Li, Yixuan
    THIRTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FOURTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE / TWELVETH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2022, : 10051 - 10059
  • [49] Provable Guarantees for Understanding Out-of-Distribution Detection
    Morteza, Peyman
    Li, Yixuan
    THIRTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FOURTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE / TWELVETH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2022, : 7831 - 7840
  • [50] Your Out-of-Distribution Detection Method is Not Robust!
    Azizmalayeri, Mohammad
    Moakhar, Arshia Soltani
    Zarei, Arman
    Zohrabi, Reihaneh
    Manzuri, Mohammad Taghi
    Rohban, Mohammad Hossein
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,