Skip to main content

2017 | OriginalPaper | Buchkapitel

Unsupervised Multi-view Subspace Learning via Maximizing Dependence

verfasst von : Meixiang Xu, Zhenfeng Zhu, Yao Zhao

Erschienen in: Computer Vision

Verlag: Springer Singapore

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config
loading …

Abstract

The recent years have witnessed the great significance of learning from multi-view data in real-world tasks, such as clustering, classification and retrieval. In this paper, we propose an unsupervised dependence (correlation) maximization model, referred to as UDM, for multi-view subspace learning. Our proposed model is based on Hilbert-Schmidt Independence Criterion (HSIC), a kernel-based technique for measuring dependence between two random variables statistically. In the proposed model, sparse constraint on the projection matrix for each view is imposed as regularizations, playing the role of feature selection, which enables to capture more discriminative subspace representations. To efficiently solve the formulated optimization problem, an iterative optimizing algorithm is designed. Experimental results on cross-modal retrieval have shown the superiority of UDM over the compared approaches and the rapid convergence speed of the optimizing algorithm.

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

  • über 102.000 Bücher
  • über 537 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Maschinenbau + Werkstoffe
  • Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 390 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Maschinenbau + Werkstoffe




 

Jetzt Wissensvorsprung sichern!

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 340 Zeitschriften

aus folgenden Fachgebieten:

  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Versicherung + Risiko




Jetzt Wissensvorsprung sichern!

Literatur
1.
Zurück zum Zitat Cichocki, A., Yang, H.H.: A new learning algorithm for blind signal separation. NIPS 3, 757–763 (1996) Cichocki, A., Yang, H.H.: A new learning algorithm for blind signal separation. NIPS 3, 757–763 (1996)
2.
Zurück zum Zitat Andrew, G., Arora, R., Blimes, J., Livescu, K.: Deep canonical correlation analysis. In: ICML, pp. 1247–1255 (2013) Andrew, G., Arora, R., Blimes, J., Livescu, K.: Deep canonical correlation analysis. In: ICML, pp. 1247–1255 (2013)
3.
Zurück zum Zitat Cao, X., Zhang, C., Fu, H., Liu, S., Zhang, H.: Diversity-induced multi-view subspace clustering. In: CVPR, pp. 586–594 (2015) Cao, X., Zhang, C., Fu, H., Liu, S., Zhang, H.: Diversity-induced multi-view subspace clustering. In: CVPR, pp. 586–594 (2015)
4.
Zurück zum Zitat Chu, D., Liao, L., Ng, M.K., Zhang, X.: Sparse canonical correlation analysis:new formulation and algorithm. TPAMI 35(12), 3050–3065 (2013)CrossRef Chu, D., Liao, L., Ng, M.K., Zhang, X.: Sparse canonical correlation analysis:new formulation and algorithm. TPAMI 35(12), 3050–3065 (2013)CrossRef
5.
Zurück zum Zitat Chua, T.-S., Tang, J., Hong, R., Li, H., Luo, Z., Zhang, Y.: Nus-wide: a real-world web image database from national university of Singapore. In: CIVR (2009) Chua, T.-S., Tang, J., Hong, R., Li, H., Luo, Z., Zhang, Y.: Nus-wide: a real-world web image database from national university of Singapore. In: CIVR (2009)
6.
Zurück zum Zitat Principe, J.C.: Information theory, machine learning, and reproducing kernel Hilbert spaces. In: Information Theoretic Learning. Information Science and Statistics, pp. 1–45. Springer, New York (2010) Principe, J.C.: Information theory, machine learning, and reproducing kernel Hilbert spaces. In: Information Theoretic Learning. Information Science and Statistics, pp. 1–45. Springer, New York (2010)
7.
Zurück zum Zitat Fang, Z., Zhang, Z.: Simultaneously combining multi-view multi-label learning with maximum margin classification. In: ICDM, pp. 864–869 (2012) Fang, Z., Zhang, Z.: Simultaneously combining multi-view multi-label learning with maximum margin classification. In: ICDM, pp. 864–869 (2012)
8.
Zurück zum Zitat Gangeh, M.J., Fewzee, P., Ghodsi, A., Kamel, M.S., Karray, F.: Kernelized supervised dictionary learning. TSP 61(19), 4753–4767 (2013)MathSciNet Gangeh, M.J., Fewzee, P., Ghodsi, A., Kamel, M.S., Karray, F.: Kernelized supervised dictionary learning. TSP 61(19), 4753–4767 (2013)MathSciNet
9.
Zurück zum Zitat Gangeh, M.J., Fewzee, P., Ghodsi, A., Kamel, M.S., Karray, F.: Multi-view supervised dictionary learning in speech emotion recognition. ACM Trans. Audio Speech Lang. Process. 22(6), 1056–1068 (2014)CrossRef Gangeh, M.J., Fewzee, P., Ghodsi, A., Kamel, M.S., Karray, F.: Multi-view supervised dictionary learning in speech emotion recognition. ACM Trans. Audio Speech Lang. Process. 22(6), 1056–1068 (2014)CrossRef
11.
Zurück zum Zitat Hotelling, H.: Relations between two sets of variates. Biometrika 28(3/4), 321–377 (1936)CrossRefMATH Hotelling, H.: Relations between two sets of variates. Biometrika 28(3/4), 321–377 (1936)CrossRefMATH
12.
Zurück zum Zitat Torkkola, K.: Feature extraction by non-parametric mutual information maximization. J. Mach. Learn. Res. 3(3), 1415–1438 (2003)MathSciNetMATH Torkkola, K.: Feature extraction by non-parametric mutual information maximization. J. Mach. Learn. Res. 3(3), 1415–1438 (2003)MathSciNetMATH
13.
Zurück zum Zitat Ngiam, J., Khosla, A., Kim, M., Nam, J., Lee, H., Ng, A.: Multimodal deep learning. In: ICML, pp. 689–696 (2011) Ngiam, J., Khosla, A., Kim, M., Nam, J., Lee, H., Ng, A.: Multimodal deep learning. In: ICML, pp. 689–696 (2011)
14.
Zurück zum Zitat Nie, F., Huang, H., Cai, X., Ding, C.: Efficient and robust feature selection via joint l2,1-norms minimization. In: NIPS, pp. 1813–1821 (2010) Nie, F., Huang, H., Cai, X., Ding, C.: Efficient and robust feature selection via joint l2,1-norms minimization. In: NIPS, pp. 1813–1821 (2010)
15.
Zurück zum Zitat Rasiwasia, N., Pereira, J.C., Coviello, E., Doyle, G., Lanckriet, G.R.G., Levy, R., Vasconcelos, N.: A new approach to cross-modal multimedia retrieval. In ICMM, pp. 251–260 (2010) Rasiwasia, N., Pereira, J.C., Coviello, E., Doyle, G., Lanckriet, G.R.G., Levy, R., Vasconcelos, N.: A new approach to cross-modal multimedia retrieval. In ICMM, pp. 251–260 (2010)
16.
Zurück zum Zitat Akaho, S.: A kernel method for canonical correlation analysis. In: IMPS 2001 (2007) Akaho, S.: A kernel method for canonical correlation analysis. In: IMPS 2001 (2007)
17.
Zurück zum Zitat Schölkopf, B., Mika, S., Smola, A., Rätsch, G., Müller, K.R.: Kernel PCA pattern reconstruction via approximate pre-images. In: Niklasson, L., Bodén, M., Ziemke, T. (eds.) ICANN 1998. Perspectives in Neural Computing, pp. 147–152. Springer, London (1998). https://doi.org/10.1007/978-1-4471-1599-1_18 Schölkopf, B., Mika, S., Smola, A., Rätsch, G., Müller, K.R.: Kernel PCA pattern reconstruction via approximate pre-images. In: Niklasson, L., Bodén, M., Ziemke, T. (eds.) ICANN 1998. Perspectives in Neural Computing, pp. 147–152. Springer, London (1998). https://​doi.​org/​10.​1007/​978-1-4471-1599-1_​18
18.
Zurück zum Zitat Sharma, A., Jacobs, D.W.: Bypassing synthesis: PLS for face recognition with pose, low-resolution and sketch. In: CVPR, pp. 593–600 (2011) Sharma, A., Jacobs, D.W.: Bypassing synthesis: PLS for face recognition with pose, low-resolution and sketch. In: CVPR, pp. 593–600 (2011)
19.
Zurück zum Zitat Tae-Kyun, K., Kittler, J., Cipolla, R.: Discriminative learning and recognition of image set classess using canonical correlation. TPAMI 29(6), 1005–1018 (2007)CrossRef Tae-Kyun, K., Kittler, J., Cipolla, R.: Discriminative learning and recognition of image set classess using canonical correlation. TPAMI 29(6), 1005–1018 (2007)CrossRef
20.
Zurück zum Zitat Tenenbaum, J.B., Freeman, W.T.: Separating style and content with bilinear models. Neural Comput. 12(6), 1247–1283 (2000)CrossRef Tenenbaum, J.B., Freeman, W.T.: Separating style and content with bilinear models. Neural Comput. 12(6), 1247–1283 (2000)CrossRef
21.
Zurück zum Zitat Wang, K., He, R., Wang, L., Wang, W., Tan, T.: Joint feature selection and subspace learning for cross-modal retrieval. TPAMI 38(10), 2010–2023 (2016)CrossRef Wang, K., He, R., Wang, L., Wang, W., Tan, T.: Joint feature selection and subspace learning for cross-modal retrieval. TPAMI 38(10), 2010–2023 (2016)CrossRef
22.
Zurück zum Zitat Wang, W., Arora, R., Livescu, K., Bilmes, J.: On deep multi-view representation learning. In: ICML (2015) Wang, W., Arora, R., Livescu, K., Bilmes, J.: On deep multi-view representation learning. In: ICML (2015)
23.
Zurück zum Zitat Wei, Y., Zhao, Y., Lu, C., Wei, S., Liu, L., Zhu, Z., Yan, S.: Cross-modal retrieval with cnn visual features: A new baseline. TCB 47(2), 449–460 (2017) Wei, Y., Zhao, Y., Lu, C., Wei, S., Liu, L., Zhu, Z., Yan, S.: Cross-modal retrieval with cnn visual features: A new baseline. TCB 47(2), 449–460 (2017)
Metadaten
Titel
Unsupervised Multi-view Subspace Learning via Maximizing Dependence
verfasst von
Meixiang Xu
Zhenfeng Zhu
Yao Zhao
Copyright-Jahr
2017
Verlag
Springer Singapore
DOI
https://doi.org/10.1007/978-981-10-7302-1_12