融合动态残差的多源域自适应算法研究

doi:10.3778/j.issn.1002-8331.2108-0185

摘要/Abstract

摘要： 多源域自适应问题通常是指拥有多个源域与单个目标域的场景。常见做法是依据域标签两两对齐源域与目标域分布，通过减小域间距离，将分布映射到共同隐空间内，去预测未知目标域的数据分类。源数据集通常需要域标签，且模型在经过训练阶段后，参数固定，这就很难达到拟合未知目标域分布的目的。基于动态残差块的多源域自适应算法不是从域的角度而是从数据自身特征映射生成神经网络参数，不需要域标签，将多源域自适应问题转化为单源域问题。而且动态残差块能够跨阶段的根据输入数据特征改变网络参数，更好地让网络参数拟合未经训练的目标域数据分布，简化了多源域自适应的模型设计复杂程度，减少了数据准备工作量。实验结果表明，在模型中引入动态残差块，与静态模型相比准确率提高了8.1%，同时也节约了模型运行的时间和空间。

关键词: 域自适应, 动态残差块, 多源域自适应, 迁移学习, 深度学习

Abstract: Multi-source domain adaptation problem usually refers to a scene with multiple source domains and a single target domain. The common approach is to align the distribution of source domain and target domain according to the domain label, and map the distribution to the common hidden space by reducing the distance between domains to predict the data classification of unknown target domain. The source data set usually needs domain label, and the parameters of the model are fixed after the training stage, which is difficult to fit the distribution of unknown target domain. The multi-source domain adaptive algorithm based on dynamic residual block generates neural network parameters not from the perspective of domain, but from the feature mapping of data itself. Without domain label, the multi-source domain adaptive problem is transformed into a single source domain problem. Moreover, the dynamic residual block can change the network parameters across stages according to the characteristics of the input data, better let the network parameters fit the untrained target domain data distribution, simplify the complexity of multi-source domain adaptive model design, and reduce the workload of data preparation. The experimental results show that the accuracy is improved by 8.1% compared with the static model, and the running time and space of the model are saved.

Key words: domain adaption, dynamic residual block, multi-source domain adaption, transfer learning, deep learning

王斌, 李昕. 融合动态残差的多源域自适应算法研究[J]. 计算机工程与应用, 2022, 58(7): 162-166.

WANG Bin, LI Xin. Research on Multi-Source Domain Adaptive Algorithm Integrating Dynamic Residuals[J]. Computer Engineering and Applications, 2022, 58(7): 162-166.

参考文献

[1] WANG M，DENG W.Deep visual domain adaptation：A survey[J].Neurocomputing，2018，312：135-153.
[2] SUN B，SAENKO K.Deep CORAL：Correlation alignment for deep domain adaptation[C]//Proceedings of European Conference on Computer Vision，2016：443-450.
[3] YANG J，YAN R，HAUPTMANN A G，et al.Cross-domain video concept detection using adaptive svms[C]//Proceedings of International Conference on Multimedia.Carnegie Mellon University，2007：188-197.
[4] BLITZER J，CRAMMER K，KULESZA A，et al.Learning Bounds for Domain Adaptation[C]//Advances in Neural Information Processing Systems，2008：129-136.
[5] XU R J，CHEN Z L，ZUO W M，et al.Deep cocktail network：Multi-source unsupervised domain adaptation with category shift[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：3964-3973.
[6] CHEN Y，DAI X，LIU M，et al.Dynamic convolution：Attention over convolution kernels[C]//Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition（CVPR），2020.
[7] CAO Y，XU J，LIN S，et al.GCNet：Non-local networks meet squeeze-excitation networks and beyond[C]//Proceedings of the IEEE International Conference on Computer Vision Workshops，2019.
[8] LIN J，RAO Y M，LU J W，et al.Runtime neural pruning[C]//Advances in Neural Information Processing Systems，2017：2181-2191.
[9] YANG B，BENDER G，LE Q V，et al.Condconv：Conditionally parameterized convolutions for efficient inference[C]//Advances in Neural Information Processing Systems，2019：1307-1318.
[10] SAITO K，WATANABE K，USHIKU Y，et al.Maximum classifier discrepancy for unsupervised domain adaptation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：3723-3732.
[11] CARIUCCI F M，PORZI L，CAPUTO B，et al.Autodial：Automatic domain alignment layers[C]//Proceedings of 2017 IEEE International Conference on Computer Vision（ICCV），2017：5077-5085.
[12] WU Z，NAGARAJAN T，KUMAR A，et al.BlockDrop：Dynamic inference paths in residual networks[C]//Proceedings of CVPR，IEEE Computer Society Conference on Computer Vision and Pattern Recognition，2017.
[13] IANDOLA F N，HAN S，MOSKEWICZ M W，et al.Squeezenet：Alexnet-level accuracy with 50x fewer parameters and <1?MB model size[J].arXiv：1602.07360，2016.
[14] YANG L Y，BALAJI Y，LIM S N，et al.Curriculum manager for source selection in multisource domain adaptation[J].arXiv：2007.01261，2020.
[15] LONG M，YUE C，CAO Z，et al.Transferable representation learning with deep adaptation networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2018，41：3071-3085.
[16] TZENG E，HOFFMAN J，SAENKO K，et al.Adversarial discriminative domain adaptation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：7167-7176.
[17] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[18] GANIN Y，LEMPITSKY V.Unsupervised domain adaptation by backpropagation[C]//Proceedings of International Conference on Machine Learning，2015：1180-1189.
[19] PENG X C，BAI Q X，XIA X D，et al.Moment matching for multi-source domain adaptation[C]//Proceedings of the IEEE International Conference on Computer Vision，2019：1406-1415.
[20] LI Y S，YUAN L，CHEN Y P，et al.Dynamic transfer for multi-source domain adaptation[C]//Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition，2021：1-10.