Γ-stochastic neighbour embedding for feed-forward data visualization

Iain Rice

doi:10.1177/1473871617715212

Γ-stochastic neighbour embedding for feed-forward data visualization

Iain Rice

College of Engineering and Physical Sciences

Research output: Contribution to journal › Article › peer-review

Abstract

t-distributed Stochastic Neighbour Embedding (t-SNE) is one of the most popular nonlinear dimension reduction techniques used in multiple application domains. In this paper we propose a variation on the embedding neighbourhood distribution, resulting in Γ-SNE, which can construct a feed-forward mapping using an RBF network. We compare the visualizations generated by Γ-SNE with those of t-SNE and provide empirical evidence suggesting the network is capable of robust interpolation and automatic weight regularization.

Original language	English
Pages (from-to)	306-315
Journal	Information Visualization
Volume	17
Issue number	4
Early online date	7 Jul 2017
DOIs	https://doi.org/10.1177/1473871617715212
Publication status	Published - 1 Oct 2018

Bibliographical note

Γ-SNE for feed-forward data visualisation
Rice, I. 7 Jul 2017 In : Information Visualization. accepted. Copyright © 2017 The Author. Reprinted by permission of SAGE Publications.

Keywords

Stochastic neighbour embedding
gamma distribution
visualization
radial basis function network
NeuroScale

Access to Document

10.1177/1473871617715212

Γ-stochastic neighbour embedding for feed-forward data visualizationAccepted author manuscript, 6.13 MB

Cite this

@article{8a273addf41f4cfeb11fe9cdc9a33107,

title = "Γ-stochastic neighbour embedding for feed-forward data visualization",

abstract = "t-distributed Stochastic Neighbour Embedding (t-SNE) is one of the most popular nonlinear dimension reduction techniques used in multiple application domains. In this paper we propose a variation on the embedding neighbourhood distribution, resulting in Γ-SNE, which can construct a feed-forward mapping using an RBF network. We compare the visualizations generated by Γ-SNE with those of t-SNE and provide empirical evidence suggesting the network is capable of robust interpolation and automatic weight regularization.",

keywords = "Stochastic neighbour embedding, gamma distribution, visualization, radial basis function network, NeuroScale",

author = "Iain Rice",

note = "Γ-SNE for feed-forward data visualisation Rice, I. 7 Jul 2017 In : Information Visualization. accepted. Copyright {\textcopyright} 2017 The Author. Reprinted by permission of SAGE Publications.",

year = "2018",

month = oct,

day = "1",

doi = "10.1177/1473871617715212",

language = "English",

volume = "17",

pages = "306--315",

journal = "Information Visualization",

issn = "1473-8724",

publisher = "SAGE",

number = "4",

}

TY - JOUR

T1 - Γ-stochastic neighbour embedding for feed-forward data visualization

AU - Rice, Iain

PY - 2018/10/1

Y1 - 2018/10/1

N2 - t-distributed Stochastic Neighbour Embedding (t-SNE) is one of the most popular nonlinear dimension reduction techniques used in multiple application domains. In this paper we propose a variation on the embedding neighbourhood distribution, resulting in Γ-SNE, which can construct a feed-forward mapping using an RBF network. We compare the visualizations generated by Γ-SNE with those of t-SNE and provide empirical evidence suggesting the network is capable of robust interpolation and automatic weight regularization.

AB - t-distributed Stochastic Neighbour Embedding (t-SNE) is one of the most popular nonlinear dimension reduction techniques used in multiple application domains. In this paper we propose a variation on the embedding neighbourhood distribution, resulting in Γ-SNE, which can construct a feed-forward mapping using an RBF network. We compare the visualizations generated by Γ-SNE with those of t-SNE and provide empirical evidence suggesting the network is capable of robust interpolation and automatic weight regularization.

KW - Stochastic neighbour embedding

KW - gamma distribution

KW - visualization

KW - radial basis function network

KW - NeuroScale

UR - https://journals.sagepub.com/doi/10.1177/1473871617715212

U2 - 10.1177/1473871617715212

DO - 10.1177/1473871617715212

M3 - Article

SN - 1473-8724

VL - 17

SP - 306

EP - 315

JO - Information Visualization

JF - Information Visualization

IS - 4

ER -

Γ-stochastic neighbour embedding for feed-forward data visualization

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this