Deep Clustering With Variational Autoencoder

An autoencoder that learns a latent space in an unsupervised manner has many applications in signal processing. However, the latent space of an autoencoder does not pursue the same clustering goal as Kmeans or GMM. A recent work proposes to artificially re-align each point in the latent space of an...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE signal processing letters Ročník 27; s. 231 - 235
Hlavní autoři: Lim, Kart-Leong, Jiang, Xudong, Yi, Chenyu
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York IEEE 2020
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:
Bayes methods
Clustering
Datasets
Energy management
Euclidean geometry
Gaussian distribution
Optimization
Photovoltaic cells
Probabilistic logic
Probabilistic models
Probability distribution
Random variables
Signal processing
Statistical analysis
ISSN:1070-9908, 1558-2361
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:An autoencoder that learns a latent space in an unsupervised manner has many applications in signal processing. However, the latent space of an autoencoder does not pursue the same clustering goal as Kmeans or GMM. A recent work proposes to artificially re-align each point in the latent space of an autoencoder to its nearest class neighbors during training (Song et al. 2013). The resulting new latent space is found to be much more suitable for clustering, since clustering information is used. Inspired by previous works (Song et al. 2013), in this letter we propose several extensions to this technique. First, we propose a probabilistic approach to generalize Song's approach, such that Euclidean distance in the latent space is now represented by KL divergence. Second, as a consequence of this generalization we can now use probability distributions as inputs rather than points in the latent space. Third, we propose using Bayesian Gaussian mixture model for clustering in the latent space. We demonstrated our proposed method on digit recognition datasets, MNIST, USPS and SHVN as well as scene datasets, Scene15 and MIT67 with interesting findings.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1070-9908
1558-2361
DOI:10.1109/LSP.2020.2965328