A Flow Artist for High-Dimensional Cellular Data
Abstract
We consider the problem of embedding point cloud data sampled from an underlying manifold with an associated flow or velocity. Such data arises in many contexts where static snapshots of dynamic entities are measured, including in high-throughput biology such as single-cell transcriptomics. Existing embedding techniques either do not utilize velocity information or embed the coordinates and velocities independently, i.e., they either impose velocities on top of an existing point embedding or embed points within a prescribed vector field. Here we present FlowArtist, a neural network that embeds points while jointly learning a vector field around the points. The combination allows FlowArtist to better separate and visualize velocity-informed structures. Our results, on toy datasets and single-cell RNA velocity data, illustrate the value of utilizing coordinate and velocity information in tandem for embedding and visualizing high-dimensional data.
Cite
@article{arxiv.2308.00176,
title = {A Flow Artist for High-Dimensional Cellular Data},
author = {Kincaid MacDonald and Dhananjay Bhaskar and Guy Thampakkul and Nhi Nguyen and Joia Zhang and Michael Perlmutter and Ian Adelstein and Smita Krishnaswamy},
journal= {arXiv preprint arXiv:2308.00176},
year = {2023}
}
Comments
Accepted for publication in 2023 IEEE 33rd International Workshop on Machine Learning for Signal Processing (MLSP)