Abstract
We present a novel approach for the fast approximation of the discrete Gauss transform in higher dimensions. The algorithm is based on the dual-tree technique and introduces a new Taylor series expansion. It compares favorably to existing methods especially when it comes to higher dimensions and a broad range of bandwidths. Numerical results with different datasets in up to 62 dimensions demonstrate its performance.
Similar content being viewed by others
Notes
We choose direct computation as the only option on leaf level, since it always outperformed Taylor approximation in our numerical experiments with our choice of the parameter q=30. Furthermore, we stop the recursion as soon as one of the nodes S or T is a leaf. In case the other node is still associated to a larger subtree, it might be favorable to further descend in this particular subtree only. However, in our test cases, the sources and targets coincide anyway. Therefore, the two trees are identical and such a general situation can not appear at all.
References
Ayyagari, V.R., Boughorbel, F., Koschan, A., Abidi, M.A.: A new method for automatic 3D face registration. In: Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., vol. 3, p. 119 (2005)
Baxter, B.J.C., Roussos, G.: A new error estimate of the fast Gauss transform. SIAM J. Sci. Comput. 24(1), 257–259 (2002)
Beatson, R., Greengard, L.: A short course on fast multipole methods. In: Wavelets, Multilevel Methods and Elliptic PDEs, pp. 1–37. Oxford University Press, London (1997)
Boughorbel, F., Koschan, A., Abidi, M.: A new multi-sensor registration technique for three-dimensional scene modeling with application to unmanned vehicle mobility enhancement. In: Unmanned Ground Veh. Technol. VII (SPIE Conf. Proc.), vol. 5804, pp. 174–181 (2005)
Broadie, M., Yamamoto, Y.: Application of the fast Gauss transform to option pricing. Manag. Sci. 49(8), 1071–1088 (2003)
Broadie, M., Yamamoto, Y.: A double-exponential fast Gauss transform algorithm for pricing discrete path-dependent options. Oper. Res. 53(5), 764–779 (2005)
Elgammal, A., Duraiswami, R., Davis, L.S.: Efficient kernel density estimation using the fast Gauss transform with applications to color modeling and tracking. IEEE Trans. Pattern Anal. Mach. Intell. 25(11), 1499–1504 (2003)
François, D., Wertz, V., Verleysen, M.: About the locality of kernels in high-dimensional spaces. In: Proc. Int. Symp. Appl. Stoch. Models Data Anal. (ASMDA), vol. 8, pp. 238–245 (2005)
Greengard, L., Rokhlin, V.: A fast algorithm for particle simulations. J. Comput. Phys. 73(2), 325–348 (1987)
Greengard, L., Strain, J.: The fast Gauss transform. SIAM J. Sci. Stat. Comput. 12(1), 79–94 (1991)
Hegland, M., Pestov, V.: Additive models in high dimensions. In: Proc. 12th Comput. Tech. Appl. Conf. (CTAC-2004). ANZIAM J., vol. 46, pp. C1205–C1221 (2005)
Keiner, J., Kunis, S., Potts, D.: Fast summation of radial functions on the sphere. Computing 78, 1–15 (2006)
Kunis, S.: Nonequispaced FFT: Generalisation and Inversion. Shaker, Aachen (2006)
Lee, D., Gray, A.: Faster Gaussian summation: theory and experiment. In: Proc. 22nd Annu. Conf. Uncertain. Artif. Intell. (UAI-06). AUAI Press, Arlington (2006)
Lee, D., Gray, A., Moore, A.: Dual-tree fast Gauss transforms. Adv. Neural Inf. Process. Syst. 18, 747–754 (2006)
Matsumoto, M., Nishimura, T.: Mersenne twister: a 623-dimensionally equidistributed uniform pseudo-random number generator. ACM Trans. Model. Comput. Simul. 8(1), 3–30 (1998)
Morariu, V.I., Srinivasan, B.V., Raykar, V.C., Duraiswami, R., Davis, L.S.: Automatic online tuning for fast Gaussian summation. In: Proc. 22nd Annu. Conf. Neural Inf. Process. Syst., pp. 1113–1120. MIT Press, Cambridge (2008)
Raykar, V.C., Duraiswami, R.: Very fast optimal bandwidth selection for univariate kernel density estimation. Tech. rep. CS-TR-4774, Dep. Comput. Sci., Univ. Maryland, College Park (2005)
Raykar, V.C., Yang, C., Duraiswami, R., Gumerov, N.A.: Fast computation of sums of Gaussians in high dimensions. Tech. rep. CS-TR-4767, Dep. Comput. Sci., Univ. Maryland, College Park (2005)
Spivak, M., Veerapaneni, S.K., Greengard, L.: The fast generalized Gauss transform. SIAM J. Sci. Comput. 32(5), 3092–3107 (2010)
Tausch, J., Weckiewicz, A.: Multidimensional fast Gauss transforms by Chebyshev expansions. SIAM J. Sci. Comput. 31, 3547–3565 (2009)
Verleysen, M.: Learning high-dimensional data. In Limit. Future Trends Neural Comput., vol. 186, pp. 141–162 (2003)
White, D.A., Jain, R.: Similarity indexing: algorithms and performance. In: Storage and Retr. for Still Image and Video Databases IV, vol. 2670, pp. 62–73 (1996)
Wissel, D.: The discrete Gauss transform—fast approximation algorithms and applications in high dimensions. Diploma thesis. Inst. Numer. Simul., Univ. Bonn (2008)
Yang, C., Duraiswami, R., Davis, L.: Efficient kernel machines using the improved fast Gauss transform. In: Adv. Neural Inf. Process. Syst., vol. 17, pp. 1561–1568. MIT Press, Cambridge (2004)
Yang, C., Duraiswami, R., Gumerov, N.A., Davis, L.: Improved fast Gauss transform and efficient kernel density estimation. In: Proc. 9th IEEE Int. Conf. Comput. Vis., vol. 1, pp. 664–671 (2003)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Griebel, M., Wissel, D. Fast Approximation of the Discrete Gauss Transform in Higher Dimensions. J Sci Comput 55, 149–172 (2013). https://doi.org/10.1007/s10915-012-9626-3
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10915-012-9626-3