Abstract:$t$-SNE is an embedding method that the data science community has widely Two interesting characteristics of t-SNE are the structure preservation property and the answer to the crowding problem, where all neighbors in high dimensional space cannot be represented correctly in low dimensional space. $t$-SNE preserves the local neighborhood, and similar items are nicely spaced by adjusting to the local density. These two characteristics produce a meaningful representation, where the cluster area is proportional to its size in number, and relationships between clusters are materialized by closeness on the embedding. This algorithm is non-parametric, therefore two initializations of the algorithm would lead to two different embedding. In a forensic approach, analysts would like to compare two or more datasets using their embedding. An approach would be to learn a parametric model over an embedding built with a subset of data. While this approach is highly scalable, points could be mapped at the same exact position, making them indistinguishable. This type of model would be unable to adapt to new outliers nor concept drift. This paper presents a methodology to reuse an embedding to create a new one, where cluster positions are preserved. The optimization process minimizes two costs, one relative to the embedding shape and the second relative to the support embedding' match. The proposed algorithm has the same complexity than the original $t$-SNE to embed new items, and a lower one when considering the embedding of a dataset sliced into sub-pieces. The method showed promising results on a real-world dataset, allowing to observe the birth, evolution and death of clusters. The proposed approach facilitates identifying significant trends and changes, which empowers the monitoring high dimensional datasets' dynamics.

Low-dimensional Data Embedding via Robust Ranking

Locality Pursuit Embedding

Entropic Optimal Transport Eigenmaps for Nonlinear Alignment and Joint Embedding of High-Dimensional Datasets

Index $t$-SNE: Tracking Dynamics of High-Dimensional Datasets with Coherent Embeddings

A New Low-Rank Structurally Incoherent Algorithm for Robust Image Feature Extraction

Efficient Algorithms for t-distributed Stochastic Neighborhood Embedding

Robust sparse low-rank embedding for image dimension reduction

Robust Low-rank Tensor Decomposition with the L2 Criterion

Deep Recursive Embedding for High-Dimensional Data

Normed Spaces for Graph Embedding

Low-rank 2D Local Discriminant Graph Embedding for Robust Image Feature Extraction

Sailing in high-dimensional spaces: Low-dimensional embeddings through angle preservation

ENS-t-SNE: Embedding Neighborhoods Simultaneously t-SNE

Dual low-rank structure embedding for robust visual information processing

Robust Discriminative t-Linear Subspace Learning for Image Feature Extraction

Statistical embedding: Beyond principal components

Word Embeddings via Tensor Factorization

Deep Randomized Ensembles for Metric Learning

Robust Neighborhood Embedding for Unsupervised Feature Selection.

Deep and Structured Robust Information Theoretic Learning for Image Analysis

Low-Rank Discriminative Embedding Regression for Robust Feature Extraction of Hyperspectral Images via Weighted Schatten p-Norm Minimization