Visualizing Large-scale and High-dimensional Data

Abstract

We study the problem of visualizing large-scale and highdimensional data in a low-dimensional (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 low-dimensional space with the structure preserved. These two steps suffer from considerable computational costs, preventing the state-ofthe-art methods such as the t-SNE from scaling to largescale and high-dimensional data (e.g., millions of data points and hundreds of dimensions). We propose the LargeVis, a technique that first constructs an accurately approximated K-nearest neighbor graph from the data and then layouts the graph in the low-dimensional space. Comparing to t-SNE, 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 real-world data sets demonstrate that the LargeVis outperforms the state-of-the-art methods in both efficiency and effectiveness. The hyper-parameters of LargeVis are also much more stable over different data sets.

Generalizable and Scalable Visualization of Single-Cell Data Using Neural Networks

“…Our approach achieves significantly better scalability than t-SNE, which is on the verge of being impractical for datasets with more than a million cells. Although a number of recent studies introduced new techniques for improving the scalability of data visualization tools (Dzwinel and Wcisło, 2015; Tang et al, 2016), they do not address the lack of generalizability that net-SNE overcomes.…”

VisAGE: Integrating external knowledge into electronic medical record visualization

“…Thus, we compared the results between using M versus using M in meaningful downstream visualization applications. Specifically, following previous studies, 17,18 we used t-distributed stochastic neighbor embedding (t-SNE), 19 an algorithm that can efficiently model high-dimensional objects as two-dimensional points, which makes it especially well-suited for visualizing our dataset. We generated our visualization by running t-SNE with default settings on the patient profile matrix M for the baseline and M for VisAGE.…”

“…However, it builds a k-NN network directly from the data, and then reduces the network to two dimensions without using external information. 18 Another study built upon LargeVis to visualize single cells, but still also directly computed embeddings from a k-NN network without utilizing external data. 17 Marlin et al visualized a pattern discovery model's clustering parameters in the context of EMR analysis.…”

CyTOF workflow: differential discovery in high-throughput high-dimensional cytometry datasets

“…To speed up the t-SNE analysis, one could use a multicore version that is available via the Rtsne.multicore package. Alternative algorithms, such as ( Tang et al , 2016) (available via the largeVis package), can be used for dimensionality reduction of very large datasets without downsampling. Alternatively, the dimensionality reduction can be performed on the codes of the SOM, at a resolution specified by the user (see Figure 12).…”