ReFine: Re-randomization before Fine-tuning for Cross-domain Few-shot Learning

被引:4
|
作者
Oh, Jaehoon [1 ]
Kim, Sungnyun [2 ]
Ho, Namgyu [2 ]
Kim, Jin-Hwa [3 ]
Song, Hwanjun [3 ]
Yun, Se-Young [2 ]
机构
[1] KAIST DS, Daejeon, South Korea
[2] KAIST AI, Seoul, South Korea
[3] NAVER AI Lab, Sungnam, South Korea
来源
PROCEEDINGS OF THE 31ST ACM INTERNATIONAL CONFERENCE ON INFORMATION AND KNOWLEDGE MANAGEMENT, CIKM 2022 | 2022年
关键词
cross-domain; few-shot; transfer learning; re-randomization;
D O I
10.1145/3511808.3557681
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Cross-domain few-shot learning (CD-FSL), where there are few target samples under extreme differences between source and target domains, has recently attracted huge attention. Recent studies on CD-FSL generally focus on transfer learning based approaches, where a neural network is pre-trained on popular labeled source domain datasets and then transferred to target domain data. Although the labeled datasets may provide suitable initial parameters for the target data, the domain difference between the source and target might hinder fine-tuning on the target domain. This paper proposes a simple yet powerful method that re-randomizes the parameters fitted on the source domain before adapting to the target data. The re-randomization resets source-specific parameters of the source pre-trained model and thus facilitates fine-tuning on the target domain, improving few-shot performance.
引用
收藏
页码:4359 / 4363
页数:5
相关论文
共 50 条
  • [21] Cross-domain mechanism for few-shot object detection on Urine Sediment Image
    Chan, Sixian
    Wu, Binghui
    Wang, Hongqiang
    Zhou, Xiaolong
    Zhang, Guodao
    Wang, Guoqing
    COMPUTERS IN BIOLOGY AND MEDICINE, 2023, 166
  • [22] An enhanced meta-learning network with sensitivity penalty for cross-domain few-shot fault diagnosis
    Mu, Mingzhe
    Jiang, Hongkai
    Jiang, Wenxin
    Dong, Yutong
    Wu, Zhenghong
    MEASUREMENT SCIENCE AND TECHNOLOGY, 2024, 35 (09)
  • [23] Transformer-Based Cross-Domain Few-Shot Learning for Hyperspectral Target Detection
    Feng, Shou
    Wang, Xueqing
    Feng, Rui
    Xiong, Fengchao
    Zhao, Chunhui
    Li, Wei
    Tao, Ran
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2025, 63
  • [24] Domain Mapping Network for Remote Sensing Cross-Domain Few-Shot Classification
    Lu, Xiaoqiang
    Gong, Tengfei
    Zheng, Xiangtao
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2024, 62 : 1 - 11
  • [25] Incremental few-shot instance segmentation without fine-tuning on novel classes
    Zhang, Luofeng
    Weng, Libo
    Zhang, Yuanming
    Gao, Fei
    COMPUTER VISION AND IMAGE UNDERSTANDING, 2025, 254
  • [26] Cross-domain few-shot semantic segmentation for the astronaut work environment
    Sun, Qingwei
    Chao, Jiangang
    Lin, Wanhong
    ADVANCES IN SPACE RESEARCH, 2024, 74 (11) : 5934 - 5949
  • [27] Spatial-Spectral-Semantic Cross-Domain Few-Shot Learning for Hyperspectral Image Classification
    Cao, Mengxin
    Zhang, Xu
    Cheng, Jinyong
    Zhao, Guixin
    Li, Wei
    Dong, Xiangjun
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2024, 62
  • [28] Convolutional Transformer-Based Few-Shot Learning for Cross-Domain Hyperspectral Image Classification
    Peng, Yishu
    Liu, Yaru
    Tu, Bing
    Zhang, Yuwen
    IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2023, 16 : 1335 - 1349
  • [29] Cross-Domain Few-Shot Learning Based on Decoupled Knowledge Distillation for Hyperspectral Image Classification
    Feng, Shou
    Zhang, Hongzhe
    Xi, Bobo
    Zhao, Chunhui
    Li, Yunsong
    Chanussot, Jocelyn
    IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2024, 62
  • [30] A novel multiple-prototype and domain adversarial network for few-shot cross-domain fault diagnosis
    Shi, Peiming
    Dai, Siyu
    Xu, Xuefang
    Han, Dongying
    MEASUREMENT SCIENCE AND TECHNOLOGY, 2025, 36 (03)