基于最优输运的迁移学习

doi:10.16451/j.cnki.issn1003-6059.201906001

Abstract
Figure/Table
References (33)
Related Citation (4)

Download: PDF (1014 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract

The goal of transfer learning is to transfer information learned from the source domain to the target domain. A transfer learning method, Optlearn, is proposed for the case of the target domain being a sub-manifold of the source domain. The source domain is weighted to make the weighted source domain and the target domain as similar as possible. The optimal transport theory is employed to minimize the difference between the weighted source domain and the target domain. Furthermore, the dual-Sinkhorn divergence is improved to suit the sub-manifold. Meanwhile, a fast computing algorithm is proposed for Optlearn. The proposed algorithm is tested on the task of pedestrian counting. Experimental results show that Optlearn obtains good counting accuracy as well as avoids the high cost of labeling data for each fixed camera.

Key words： Transfer Learning Optimal Transport Pedestrian Counting Sub Manifold

Received: 12 April 2019

ZTFLH:

TP 181

About author:: (CHE Lingfu, master student. His research interests include machine learning, computer vision, intelligent transportation system and crowd counting.)(TIAN Yukun, master student. His research interests include machine learning, computer vision, intelligent transportation system and crowd counting.)(ZHU Haiping, Ph.D. candidate. His research interests include machine learning and computer vision.)(ZHANG Junping(Corresponding author), Ph.D., professor. His research interests include machine learning, intelligent transportation system and image processing.)

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	CHE Lingfu
	TIAN Yukun
	ZHU Haiping

Cite this article:

CHE Lingfu,TIAN Yukun,ZHU Haiping. Optimal Transport Based Transfer Learning[J]. , 2019, 32(6): 481-493.

URL:

http://manu46.magtech.com.cn/Jweb_prai/EN/10.16451/j.cnki.issn1003-6059.201906001 OR http://manu46.magtech.com.cn/Jweb_prai/EN/Y2019/V32/I6/481

[1] FERNANDO B, HABRARD A, SEBBAN M, et al. Unsupervised Visual Domain Adaptation Using Subspace Alignment // Proc of the IEEE International Conference on Computer Vision. Washington, USA: IEEE, 2013: 2960-2967.
[2] ALJUNDI R, EMONET R, MUSELET D, et al. Landmarks-Based Kernelized Subspace Alignment for Unsupervised Domain Adaptation // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2015: 56-63.
[3] LU H, ZHANG L, CAO Z G, et al. When Unsupervised Domain Adaptation Meets Tensor Representations // Proc of the IEEE International Conference on Computer Vision. Washington, USA: IEEE, 2017: 599-608.
[4] 张倩,李明,王雪松,等.一种面向多源领域的实例迁移学习.自动化学报, 2014, 40(6): 1176-1183.
(ZHANG Q, LI M, WANG X S, et al. Instance-Based Transfer Learning for Multi-source Domains. Acta Automatica Sinica, 2014, 40(6): 1176-1183.)
[5] 张景祥,王士同,邓赵红,等.融合异构特征的子空间迁移学习算法.自动化学报, 2014, 40(2): 236-246.
(ZHANG J X, WANG S T, DENG Z H, et al. A Subspace Transfer Learning Algorithm Integrating Heterogeneous Features. Acta Automatica Sinica, 2014, 40(2): 236-246.)
[6] SUGIYAMA M, NAKAJIMA S, KASHIMA H, et al. Direct Importance Estimation with Model Selection and Its Application to Covariate Shift Adaptation // SCHÖLKOPF B, PLATT J, HOFMANN T, eds. Advances in Neural Information Processing Systems 20. Cambridge, USA: The MIT Press, 2007: 1433-1440.
[7] SUN B, FENG J S, SAENKO K. Return of Frustratingly Easy Domain Adaptation[C/OL]. [2019-02-26]. https://arxiv.org/pdf/1511.05547.pdf.
[8] GONG B Q, SHI Y, SHA F, et al. Geodesic Flow Kernel for Unsupervised Domain Adaptation // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2012: 2066-2073.
[9] COURTY N, FLAMARY R, TUIA D. Domain Adaptation with Re-gularized Optimal Transport // Proc of the Joint European Confe-rence on Machine Learning and Knowledge Discovery in Databases. Berlin, Germany: Springer, 2014: 274-289
[10] TAN B, SONG Y Q, ZHONG E H, et al. Transitive Transfer Learning // Proc of the 21th ACM SIGKDD International Confe-rence on Knowledge Discovery and Data Mining. New York, USA: ACM, 2015: 1155-1164.
[11] TAN B, ZHANG Y, PAN S J, et al. Distant Domain Transfer Learning // Proc of the 31st AAAI Conference on Artificial Intelligence. Palo Alto, USA: AAAI Press, 2017: 2604-2610.
[12] ARJOVSKY M, CHINTALA S, BOTTOU L. Wasserstein Generative Adversarial Networks // Proc of the 34th International Confe-rence on Machine Learning. New York, USA: ACM, 2017: 214-223.
[13] GULRAJANI I, AHMED F, ARJOVSKY M, et al. Improved Training of Wasserstein Gans[C/OL]. [2019-02-26]. https://arxiv.org/pdf/1704.00028.pdf.
[14] CUTURI M. Sinkhorn Distances: Lightspeed Computation of Optimal Transport // BURGES C J C, BOTTOU L, WELLING M, et al., eds. Advances in Neural Information Processing Systems 26. Cambridge, USA: The MIT Press, 2013: 2292-2300.
[15] BLONDEL M, SEGUY V, ROLET A. Smooth and Sparse Optimal Transport[C/OL]. [2019-02-26]. https://arxiv.org/pdf/1710.06276.pdf.
[16] ZHAO P, ZHOU Z H. Label Distribution Learning by Optimal Transport[C/OL]. [2019-02-26]. https://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/aaai18ladot.pdf.
[17] SANJABI M, BA J, RAZAVIYAYN M, et al. On the Convergence and Robustness of Training GANs with Regularized Optimal Transport // BENGIO S, WALLACH H M, LAROCHELLE H, et al., eds. Advances in Neural Information Processing Systems 31. Cambridge, USA: The MIT Press, 2018: 7091-7101.
[18] ZHAO T, NEVATIA R, WU B. Segmentation and Tracking of Multiple Humans in Crowded Environments. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008, 30(7): 1198-1211.
[19] DOLLAR P, WOJEK C, SCHIELE B, et al. Pedestrian Detection: An Evaluation of the State of the Art. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(4): 743-761.
[20] GE W, COLLINS R T. Marked Point Processes for Crowd Counting // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2009: 2913-2920.
[21] CHAN A B, LIANG Z S J, VASCONCELOS N. Privacy Preserving Crowd Monitoring: Counting People without People Models or Tracking // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2008. DOI: 10.1109/CVPR.2008.4587569.
[22] CHAN A B, VASCONCELOS N. Bayesian Poisson Regression for Crowd Counting // Proc of the 12th IEEE International Conference on Computer Vision. Washington, USA: IEEE, 2009: 545-551.
[23] CHAN A B, VASCONCELOS N. Counting People with Low-Level Features and Bayesian Regression. IEEE Transactions on Image Processing, 2012, 21(4): 2160-2177.
[24] CHEN K, GONG S G, XIANG T, et al. Cumulative Attribute Space for Age and Crowd Density Estimation // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2013: 2467-2474.
[25] TAN B, ZHANG J P, WANG L. Semi-supervised Elastic Net for Pedestrian Counting. Pattern Recognition, 2011, 44(10/11): 2297-2304.
[26] XIA W, ZHANG J P, KRUGER U. Semisupervised Pedestrian Counting with Temporal and Spatial Consistencies. IEEE Transactions on Intelligent Transportation Systems, 2015, 16(4): 1705-1715.
[27] ZHOU Q, ZHANG J P, CHE L F, et al. Crowd Counting with Limited Labeling through Submodular Frame Selection. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(5): 1728-1738.
[28] 覃勋辉,王修飞,周曦,等.多种人群密度场景下的人群计数.中国图象图形学报, 2013, 18(4): 392-398.
(QIN X H, WANG X F, ZHOU X, et al. Counting People in Various Crowed Density Scenes Using Support Vector Regression. Journal of Image and Graphics, 2013, 18(4): 392-398.)
[29] SINDAGI V A, PATEL V M. Generating High-Quality Crowd Density Maps Using Contextual Pyramid CNNs // Proc of the IEEE International Conference on Computer Vision. Washington, USA: IEEE, 2017, I: 1879-1888.
[30] SAM D B, SURYA S, BABU R V. Switching Convolutional Neural Network for Crowd Counting // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2017: 4031-4039.
[31] ZHANG Y Y, ZHOU D S, CHEN S Q, et al. Single-Image Crowd Counting via Multi-column Convolutional Neural Network // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2016: 589-597.
[32] KANG D, DHAR D, CHAN A B. Crowd Counting by Adapting Convolutional Neural Networks with Side Information[C/OL]. [2019-02-26]. https://arxiv.org/pdf/1611.06748.pdf.
[33] 时增林,叶阳东,吴云鹏,等.基于序的空间金字塔池化网络的人群计数方法.自动化学报, 2016, 42(6): 866-874.
(SHI Z L, YE Y D, WU Y P, et al. Crowd Counting Using Rank-based Spatial Pyramid Pooling Network. Acta Automatica Sinica, 2016, 42(6): 866-874.)