Manifolk: A 3D t-SNE Visualizer

被引：0

作者：

Kodur, Krishna Chaitanya ^{[1
]}

Babu, Ashwin Ramesh ^{[1
]}

Makedon, Fillia ^{[1
]}

机构：

[1] Univ Texas Arlington, Arlington, TX 76019 USA

来源：

THE 14TH ACM INTERNATIONAL CONFERENCE ON PERVASIVE TECHNOLOGIES RELATED TO ASSISTIVE ENVIRONMENTS, PETRA 2021 | 2021年

关键词：

Deep Neural Networks; t-SNE; Clustering;

D O I：

10.1145/3453892.3466619

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

Manifolk is a tool to visualize the output of dimensionality reduction algorithms like t-SNE, PCA etc. One of this tool's main uses is that it de-clutters graphs by plotting data points pertaining to a subset of labels. The subset of labels to be plotted can be selected using the provided checkboxes. A case study on data from a publicly available action recognition dataset like UCF101 shows how this tool can help find outliers. With the rise in self-supervised methods for training deep neural networks, this tool helps researchers better visualize the embeddings learned by the model.

引用

页码：107 / 108

页数：2

共 8 条

[1] [Anonymous], Plotly Dash T-SNE on MNIST Data
[2] [Anonymous], Embeddings Projector by Tensorboard
[3] [Anonymous], TSNE GUI Github
[4] He KM, 2015, Arxiv, DOI [arXiv:1512.03385, DOI 10.48550/ARXIV.1512.03385]
[5] A Survey on Contrastive Self-Supervised Learning
Jaiswal, Ashish
Babu, Ashwin Ramesh
Zadeh, Mohammad Zaki
Banerjee, Debapriya
Makedon, Fillia
[J]. TECHNOLOGIES, 2021, 9 (01)
[6] McInnes L, 2020, Arxiv, DOI [arXiv:1802.03426, 10.21105/joss.00861]
[7] Soomro K, 2012, Arxiv, DOI [arXiv:1212.0402, DOI 10.48550/ARXIV.1212.0402]
[8] van der Maaten L, 2008, J MACH LEARN RES, V9, P2579

← 1 →