Boosting the Generalization Capability in Cross-Domain Few-shot Learning via Noise-enhanced Supervised Autoencoder

被引:43
|
作者
Liang, Hanwen [1 ]
Zhang, Qiong [2 ]
Dai, Peng [1 ]
Lu, Juwei [1 ]
机构
[1] Huawei Noahs Ark Lab, Markham, PQ, Canada
[2] Univ British Columbia, Dept Stat, Vancouver, BC, Canada
来源
2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021) | 2021年
关键词
D O I
10.1109/ICCV48922.2021.00929
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
State of the art (SOTA) few-shot learning (FSL) methods suffer significant performance drop in the presence of domain differences between source and target datasets. The strong discrimination ability on the source dataset does not necessarily translate to high classification accuracy on the target dataset. In this work, we address this cross-domain few-shot learning (CDFSL) problem by boosting the generalization capability of the model. Specifically, we teach the model to capture broader variations of the feature distributions with a novel noise-enhanced supervised autoencoder (NSAE). NSAE trains the model by jointly reconstructing inputs and predicting the labels of inputs as well as their reconstructed pairs. Theoretical analysis based on intra-class correlation (ICC) shows that the feature embeddings learned from NSAE have stronger discrimination and generalization abilities in the target domain. We also take advantage of NSAE structure and propose a two-step fine-tuning procedure that achieves better adaption and improves classification performance in the target domain. Extensive experiments and ablation studies are conducted to demonstrate the effectiveness of the proposed method. Experimental results show that our proposed method consistently outperforms SOTA methods under various conditions.
引用
收藏
页码:9404 / 9414
页数:11
相关论文
共 50 条
  • [1] DisRot: boosting the generalization capability of few-shot learning via knowledge distillation and self-supervised learning
    Ma, Chenyu
    Jia, Jinfang
    Huang, Jianqiang
    Wu, Li
    Wang, Xiaoying
    MACHINE VISION AND APPLICATIONS, 2024, 35 (03)
  • [2] Understanding Cross-Domain Few-Shot Learning Based on Domain Similarity and Few-Shot Difficulty
    Oh, Jaehoon
    Kim, Sungnyun
    Ho, Namgyu
    Kim, Jin-Hwa
    Song, Hwanjun
    Yun, Se-Young
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
  • [3] Cross-domain few-shot learning via adaptive transformer networks
    Paeedeh, Naeem
    Pratama, Mahardhika
    Ma'sum, Muhammad Anwar
    Mayer, Wolfgang
    Cao, Zehong
    Kowlczyk, Ryszard
    KNOWLEDGE-BASED SYSTEMS, 2024, 288
  • [4] Cross-domain few-shot learning via adaptive transformer networks
    Paeedeh, Naeem
    Pratama, Mahardhika
    Ma'sum, Muhammad Anwar
    Mayer, Wolfgang
    Cao, Zehong
    Kowlczyk, Ryszard
    Knowledge-Based Systems, 2024, 288
  • [5] Knowledge transduction for cross-domain few-shot learning
    Li, Pengfang
    Liu, Fang
    Jiao, Licheng
    Li, Shuo
    Li, Lingling
    Liu, Xu
    Huang, Xinyan
    PATTERN RECOGNITION, 2023, 141
  • [6] Prototypical Cross-domain Self-supervised Learning for Few-shot Unsupervised Domain Adaptation
    Yue, Xiangyu
    Zheng, Zangwei
    Zhang, Shanghang
    Gao, Yang
    Darrell, Trevor
    Keutzer, Kurt
    Vincentelli, Alberto Sangiovanni
    2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 13829 - 13839
  • [7] Cross-Domain Few-Shot Relation Extraction via Representation Learning and Domain Adaptation
    Yuan, Zhongju
    Wang, Zhenkun
    Li, Genghui
    2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN, 2023,
  • [8] Few-shot Heterogeneous Graph Learning via Cross-domain Knowledge Transfer
    Zhang, Qiannan
    Wu, Xiaodong
    Yang, Qiang
    Zhang, Chuxu
    Zhang, Xiangliang
    PROCEEDINGS OF THE 28TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, KDD 2022, 2022, : 2450 - 2460
  • [9] Few-shot Image Generation via Cross-domain Correspondence
    Ojha, Utkarsh
    Li, Yijun
    Lu, Jingwan
    Efros, Alexei A.
    Lee, Yong Jae
    Shechtman, Eli
    Zhang, Richard
    2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 10738 - 10747
  • [10] Feature extractor stacking for cross-domain few-shot learning
    Wang, Hongyu
    Frank, Eibe
    Pfahringer, Bernhard
    Mayo, Michael
    Holmes, Geoffrey
    MACHINE LEARNING, 2024, 113 (01) : 121 - 158