Efficient Algorithms for t-distributed Stochastic Neighborhood Embedding

Abstract

t-distributed Stochastic Neighborhood Embedding (t-SNE) is a method for dimensionality reduction and visualization that has become widely popular in recent years. Efficient implementations of t-SNE are available, but they scale poorly to datasets with hundreds of thousands to millions of high dimensional data-points. We present Fast Fourier Transform-accelerated Interpolation-based t-SNE (FIt-SNE), which dramatically accelerates the computation of t-SNE. The most time-consuming step of t-SNE is a convolution that we accelerate by interpolating onto an equispaced grid and subsequently using the fast Fourier transform to perform the convolution. We also optimize the computation of input similarities in high dimensions using multi-threaded approximate nearest neighbors. We further present a modification to t-SNE called "late exaggeration," which allows for easier identification of clusters in t-SNE embeddings. Finally, for datasets that cannot be loaded into the memory, we present out-of-core randomized principal component analysis (oocPCA), so that the top principal components of a dataset can be computed without ever fully loading the matrix, hence allowing for t-SNE of large datasets to be computed on resource-limited machines.

References in corpus (1)

• heatmaply: an R package for creating interactive cluster heatmaps for online publishing

Cited by in corpus (32)

• Understanding How Dimension Reduction Tools Work: An Empirical Approach to Deciphering t-SNE, UMAP, TriMAP, and PaCMAP for Data Visualization
• t-viSNE: Interactive Assessment and Interpretation of t-SNE Projections
• GraphIQA: Learning Distortion Graph Representations for Blind Image Quality Assessment
• Theoretical Foundations of t-SNE for Visualizing High-Dimensional Clustered Data
• Uniform Manifold Approximation with Two-phase Optimization
• Manifold Learning in Atomistic Simulations: A Conceptual Review
• Self-generated persistent random forces drive phase separation in growing tumors
• Heavy-tailed kernels reveal a finer cluster structure in t-SNE visualisations
• Deep Recursive Embedding for High-Dimensional Data
• Force-Directed Graph Layouts Revisited: A New Force Based on the T-Distribution
• Review of Deep Representation Learning Techniques for Brain-Computer Interfaces and Recommendations
• IAN: Iterated Adaptive Neighborhoods for manifold learning and dimensionality estimation
• Review of Single-cell RNA-seq Data Clustering for Cell Type Identification and Characterization
• Improving the Effectiveness and Efficiency of Stochastic Neighbour Embedding with Isolation Kernel
• Are Graph Neural Networks Miscalibrated?
• Tree-SNE: Hierarchical Clustering and Visualization Using t-SNE
• Visualising the Evolution of English Covid-19 Cases with Topological Data Analysis Ball Mapper
• Unsupervised learning for variability detection with Gaia DR3 photometry. The main sequence-white dwarf valley
• Interpretable Visualizations with Differentiating Embedding Networks
• A Probabilistic Graph Coupling View of Dimension Reduction
• Code-Mixing on Sesame Street: Dawn of the Adversarial Polyglots
• Revised Conditional t-SNE: Looking Beyond the Nearest Neighbors
• Deep learning tackles single-cell analysis A survey of deep learning for scRNA-seq analysis
• Breaking the curse of dimensionality with Isolation Kernel
• Integer linear programming for unsupervised training set selection in molecular machine learning
• Facilitate the Parametric Dimension Reduction by Gradient Clipping
• Factoring out prior knowledge from low-dimensional embeddings
• UMAP-assisted $K$-means clustering of large-scale SARS-CoV-2 mutation datasets
• Angular Embedding: A New Angular Robust Principal Component Analysis
• EmbedOR: Provable Cluster-Preserving Visualizations with Curvature-Based Stochastic Neighbor Embeddings
• Spaceland Embedding of Sparse Stochastic Graphs
• Towards a comprehensive visualization of structure in data