Unsupervised Deep Embedding for Clustering Analysis

19 Nov 2015  ·  Junyuan Xie, Ross Girshick, Ali Farhadi ·

Clustering is central to many data-driven application domains and has been studied extensively in terms of distance functions and grouping algorithms. Relatively little work has focused on learning representations for clustering. In this paper, we propose Deep Embedded Clustering (DEC), a method that simultaneously learns feature representations and cluster assignments using deep neural networks. DEC learns a mapping from the data space to a lower-dimensional feature space in which it iteratively optimizes a clustering objective. Our experimental evaluations on image and text corpora show significant improvement over state-of-the-art methods.

PDF Abstract

Results from the Paper


Ranked #4 on Unsupervised Image Classification on SVHN (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Image Clustering CIFAR-10 DEC Accuracy 0.301 # 31
NMI 0.25 # 27
Train set Train+Test # 1
ARI 0.161 # 28
Backbone Custom # 1
Image Clustering ImageNet-10 DEC Accuracy 0.381 # 14
NMI 0.282 # 14
Image Clustering STL-10 DEC Accuracy 0.359 # 25
NMI 0.276 # 22
Train Split Train+Test # 1
Unsupervised Image Classification SVHN DEC Acc 11.90 # 4
# of clusters (k) 10 # 1

Results from Other Papers


Task Dataset Model Metric Name Metric Value Rank Uses Extra
Training Data
Source Paper Compare
Image Clustering CIFAR-100 DEC Accuracy 0.185 # 24
NMI 0.136 # 21
Train Set Train+Test # 1
Image Clustering Imagenet-dog-15 DEC Accuracy 0.195 # 16
NMI 0.122 # 16
Image Clustering Tiny-ImageNet DEC Accuracy 0.037 # 11
NMI 0.115 # 11
Image Clustering CMU-PIE DEC (KL based) NMI 0.924 # 4
Accuracy 0.801 # 3
Image Clustering YouTube Faces DB DEC (KL based) NMI 0.446 # 4
Accuracy 0.371 # 3

Methods


No methods listed for this paper. Add relevant methods here