Advertisement

Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

Applying Ricci flow to high dimensional manifold learning

Abstract

In machine learning, a high dimensional data set such as the digital image of a human face is often viewed as a point set distributed on a differentiable manifold. In many cases, the intrinsic dimension of this manifold is low but the representation dimension of the data points is high. To ease data processing requirements, manifold learning (ML) techniques can be used to reduce a high dimensional manifold (HDM) to a low dimensional one while keeping the essential geometric properties, such as relative distances between points, unchanged. Traditional ML algorithms often assume that the local neighborhood of any point on an HDM is roughly equal to the tangent space at that point. This assumption leads to the disadvantage that the neighborhoods of points on the manifold, though they have a very different curvature, will be treated equally and will be projected to a lower dimensional space. The curvature is a different way of manifold processing, where traditional dimension reduction is ineffective at preserving the neighborhood. To overcome this obstacle, we perform an “operation” on the HDM using Ricci flow before a manifold’s dimension reduction. More precisely, with the Ricci flow, we transform each local neighborhood of the HDM to a constant curvature patch. The HDM, as a whole, is then transformed into a subset of a sphere with constant positive curvature. We compare the proposed algorithm with other traditional manifold learning algorithms. Experimental results have shown that the proposed method outperforms other ML algorithms with a better neighborhood preserving rate.

This is a preview of subscription content, log in to check access.

References

  1. 1

    Roweis S T, Saul L K. Nonlinear dimensionality reduction by locally linear embedding. Science, 2000, 290: 2323–2326

  2. 2

    Tenenbaum J B, Silva V D, Langford J C. A global geometric framework for nonlinear dimensionality reduction. Science, 2000, 290: 2319–2323

  3. 3

    Lin B B, He X F, Ye J P. A geometric viewpoint of manifold learning. Appl Inform, 2015, 2: 3

  4. 4

    Zhu T, Shen F R, Zhao J X, et al. Topology learning embedding: a fast and incremental method for manifold learning. In: Proceedings of International Conference on Neural Information Processing, 2007. 43–52

  5. 5

    Belkin M, Niyogi P. Laplacian eigenmaps and spectral technique for embedding and clustering. Neural Inform Process Syst, 2001, 14: 585–591

  6. 6

    He X F, Partha N. Locality preserving projections. In: Proceedings of International Conference Advances in Neural Information Processing Systems, 2003

  7. 7

    Li Y Y, Lu R Q. Locality preserving projection on SPD matrix Lie group: algorithm and analysis. Sci China Inf Sci, 2018, 61: 092104

  8. 8

    Zhang Z Y, Zha H B. Principal manifolds and nonlinear dimension reduction via local tangent space alignment. SIAM J Sci Comput, 2005, 26: 313–338

  9. 9

    Donoho D L, Grimes C. Hessian eigenmaps: locally linear embedding techniques for high-dimensional data. Proc Natl Acad Sci USA, 2003, 100: 5591–5596

  10. 10

    Cox T F, Cox M A A. Multidimensional Scaling. Boca Raton: CRC Press, 1990

  11. 11

    Jolliffe I T. Principal Component Analysis. Berlin: Springer, 1989

  12. 12

    Hettiarachchi R, Peters J F. Multi-manifold LLE learning in pattern recognition. Pattern Recogn, 2015, 48: 2947–2960

  13. 13

    Mahapatra S, Chandola V. S-Isomap++: multi manifold learning from streaming data. In: Proceedings of IEEE International Conference on Big Data, 2017

  14. 14

    Cao H D, Zhu X P. A complete proof of the Poincaré and geometrization conjectures-application of the Hamilton-Perelman theory of the Ricci flow. Asian J Math, 2006, 10: 165–492

  15. 15

    Kwang I K, James T, Christian T. Curvature-aware regularization on Riemannian submanifolds. In: Proceedings of IEEE International Conference on Computer Vision, 2013

  16. 16

    Xu W P, Hancock E R, Wilson R C. Ricci flow embedding for rectifying non-Euclidean dissimilarity data. Pattern Recogn, 2014, 47: 3709–3725

  17. 17

    Xu W P, Hancock E R, Wilson R C. Regularising the Ricci flow embedding. In: Proceedings of Joint IAPR International Workshops on Statistical Techniques in Pattern Recognition (SPR) and Structural and Syntactic Pattern Recognition (SSPR), 2010. 579–588

  18. 18

    Xu W P, Hancock E R, Wilson R C. Rectifying non-Euclidean similarity data using Ricci flow embedding. In: Proceedings of the 20th International Conference on Pattern Recognition, 2010

  19. 19

    Zeng W, Dimitris S, Gu D. Ricci flow for 3D shape analysis. IEEE Trans Pattern Anal Mach Intell, 2010, 32: 662–677

  20. 20

    Zeng W, Yin X T, Zeng Y, et al. 3D face matching and registration based on hyperbolic Ricci flow. In: Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, 2008

  21. 21

    Peter P. Riemannian Geometry. Berlin: Springer, 1998

  22. 22

    John M. Riemannian Manifolds. Berlin: Springer, 1997

  23. 23

    Cao H D, Chow B. Recent developments on the Ricci flow. Bull Amer Math Soc, 1999, 36: 59–75

  24. 24

    Simon B, Richard S. Riemannian manifolds of positive curvature. In: Proceedings of the International Congress of Mathematicians, 2000

  25. 25

    Brendle S, Schoen R M. Classification of manifolds with weakly 1/4-pinched curvatures. Acta Math, 2008, 200: 1–13

  26. 26

    Gorelick L, Blank M, Shechtman E, et al. Actions as space-time shapes. IEEE Trans Pattern Anal Mach Intell, 2007, 29: 2247–2253

  27. 27

    Wittman T. Mani Matlab demo. 2005. http://www.math.umn.edu/~wittman/mani/

  28. 28

    Sanguinetti G. Dimensionality reduction of clustered data sets. IEEE Trans Pattern Anal Mach Intell, 2008, 30: 535–540

Download references

Acknowledgements

This work was supported by National Key Research and Development Program of China (Grant No. 2016YFB1000902), National Natural Science Foundation of China (Grant Nos. 61472412, 61621003), Beijing Science and Technology Project, and Tsinghua-Tencent-AMSS-Joint Project.

Author information

Correspondence to Yangyang Li.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Li, Y., Lu, R. Applying Ricci flow to high dimensional manifold learning. Sci. China Inf. Sci. 62, 192101 (2019). https://doi.org/10.1007/s11432-018-9702-7

Download citation

Keywords

  • manifold learning
  • Ricci flow
  • Ricci curvature
  • dimension reduction
  • curvature estimation