swMATH ID: 
34905

Software Authors: 
Jian Tang, Jingzhou Liu, Ming Zhang, Qiaozhu Mei

Description: 
LargeVis: Visualizing Largescale and Highdimensional Data. We study the problem of visualizing largescale and highdimensional data in a lowdimensional (typically 2D or 3D) space. Much success has been reported recently by techniques that first compute a similarity structure of the data points and then project them into a lowdimensional space with the structure preserved. These two steps suffer from considerable computational costs, preventing the stateoftheart methods such as the tSNE from scaling to largescale and highdimensional data (e.g., millions of data points and hundreds of dimensions). We propose the LargeVis, a technique that first constructs an accurately approximated Knearest neighbor graph from the data and then layouts the graph in the lowdimensional space. Comparing to tSNE, LargeVis significantly reduces the computational cost of the graph construction step and employs a principled probabilistic model for the visualization step, the objective of which can be effectively optimized through asynchronous stochastic gradient descent with a linear time complexity. The whole procedure thus easily scales to millions of highdimensional data points. Experimental results on realworld data sets demonstrate that the LargeVis outperforms the stateoftheart methods in both efficiency and effectiveness. The hyperparameters of LargeVis are also much more stable over different data sets. 
Homepage: 
https://arxiv.org/abs/1602.00370

Source Code: 
https://github.com/lferry007/LargeVis

Related Software: 
largeVis;
UMAP;
tSNE;
Scikit;
UCIml;
openTSNE;
TriMap;
MNIST;
FItSNE;
word2vec;
DeepWalk;
node2vec;
PPINN;
GOSDT;
InfoGAN;
CLEVR;
IMLI;
BEAMES;
EntropySGD;
MLIC

Cited in: 
5 Publications
