Accelerating t-SNE using Tree-Based Algorithms

被引:0
作者
van der Maaten, Laurens [1 ]
机构
[1] Delft Univ Technol, Pattern Recognit & Bioinformat Grp, NL-2628 CD Delft, Netherlands
关键词
embedding; multidimensional scaling; t-SNE; space-partitioning trees; Barnes-Hut algorithm; dual-tree algorithm; NONLINEAR DIMENSIONALITY REDUCTION; ERROR ESTIMATE; OBJECT; CODE;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The paper investigates the acceleration of t-SNE an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots using two treebased algorithms. In particular, the paper develops variants of the Barnes-Hut algorithm and of the dual-tree algorithm that approximate the gradient used for learning t-SNE embeddings in 0(N log N). Our experiments show that the resulting algorithms substantially accelerate t-SNE, and that they make it possible to learn embeddings of data sets with millions of objects. Somewhat counterintuitively, the Barnes-Hut variant of t-SNE appears to outperform the dual-tree variant.
引用
收藏
页码:3221 / 3245
页数:25
相关论文
共 67 条
  • [1] [Anonymous], 2012, P ICML
  • [2] [Anonymous], P 30 S THEOR COMP
  • [3] [Anonymous], LNCS
  • [4] [Anonymous], 2013, P INT C LEARN REPR
  • [5] [Anonymous], P IEEE C COMP VIS PA
  • [6] [Anonymous], 2013, Caffe: An Open Source Convolutional Architecture for Fast Feature Embedding
  • [7] [Anonymous], 2006, Mathematica journal, DOI DOI 10.3402/QHW.V6I2.5918
  • [8] [Anonymous], 2009, TICCTR2009005
  • [9] A HIERARCHICAL O(N-LOG-N) FORCE-CALCULATION ALGORITHM
    BARNES, J
    HUT, P
    [J]. NATURE, 1986, 324 (6096) : 446 - 449
  • [10] A new error estimate of the fast Gauss transform
    Baxter, BJC
    Roussos, G
    [J]. SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2002, 24 (01) : 257 - 259