Adversarial Autoencoders

By Alireza Makhzani, Jonathon Shlens, Navdeep Jaitly, Ian J. Goodfellow. 2015

2,388 citations recorded by the source corpus.

In this paper, we propose the"adversarial autoencoder"(AAE), which is a probabilistic autoencoder that uses the recently proposed generative adversarial networks (GAN) to perform variational inference by matching the aggregated posterior of the hidden code vector of the autoencoder with an arbitrary prior distribution. Matching the aggregated posterior to the prior ensures that generating from any part of prior space results in meaningful samples. As a result, the decoder of the adversarial autoencoder learns a deep generative model that maps the imposed prior to the data distribution. We show how the adversarial autoencoder can be used in applications such as semi-supervised classification, disentangling style and content of images, unsupervised clustering, dimensionality reduction and data visualization. We performed experiments on MNIST, Street View House Numbers and Toronto Face datasets and show that adversarial autoencoders achieve competitive results in generative modeling and semi-supervised classification tasks.

Subfields: Computer Vision and Pattern Recognition, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Generative Models

Citations in the corpus

InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets (2016, background): “Similarly, VAEs [5] and Adversarial Autoencoders [9] were shown to learn representations in which class label is separated from other variations.”
Evaluating Lossy Compression Rates of Deep Generative Models (2020, methodology): “9, shows the BDMC gaps at different compression rates for the VAE, GAN and AAE experiments on the MNIST dataset.”
D2C: Diffusion-Decoding Models for Few-Shot Conditional Generation (2021, methodology): “Addressing posterior mismatch in VAEs Most methods address this mismatch problem by improving inference models [61, 48, 86], prior models [87, 3, 84], or objective functions [99, 100, 101, 1, 56]; all these approaches optimize the posterior model to be close to the prior.”
Diagnosing and Enhancing VAE Models (2019, background): “But if the measure µ gt on χ has not been accurately estimated, then where q φ ( z ) is sometimes referred to as the aggregated posterior (Makhzani et al., 2016).”
GAN Inversion: A Survey (2021, neutral): “[175] use adversarial autoencoder [181] to learn invariant features for domain generalization.”
Large Scale Adversarial Representation Learning (2019, methodology): “In the adversarial space, adversarial autoencoders [24] proposed an autoencoder-style encoder-decoder pair trained with pixel-level reconstruction cost, replacing the KL-divergence regularization of the prior used in VAEs [19] with a discriminator.”
Learning deep representations by mutual information estimation and maximization (2018, methodology): “This is done (see Figure 6 (ˆ ω, ˆ ψ ) P = arg min (cid:98) This approach is similar to what is done in adversarial autoencoders (AAE, Makhzani et al., 2015), but without a generator.”
Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks (2017, background): “Such a setup can also be seen as a special case of “adversarial autoencoders” [30], which use an adversarial loss to train the bottleneck layer of an autoencoder to match an arbitrary target distribution.”
Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks (2017, background): “While there were some attempts at combining VAEs and GANs (Makhzani et al., 2015; Larsen et al., 2015), most of these attempts are not motivated from a maximum-likelihood point of view and therefore usually do not lead to maximum-likelihood assignments.”
f-GAN: Training Generative Neural Samplers using Variational Divergence Minimization (2016, methodology): “The advantage of VAEs is in the encoder model which allows efficient inference from observation to latent representation and overall they are a compelling alternative to f -GANs and recent work has studied combinations of the two approaches [23] As an alternative to the GAN training objective the wo”
Adversarially Learned Inference (2016, methodology): “The adversarial autoencoder model (Makhzani et al., 2015) replaces the KL-divergence term with a discriminator that is trained to distinguish between approximate posterior and prior samples, which provides a more flexible approach to matching the marginal q(z) and the prior.”
Stacked Generative Adversarial Networks (2016, background): “The expressive power of deep neural networks makes them natural candidates for generative models, and several recent works have shown promising results [28, 17, 44, 36, 68, 38, 9].”
Multimodal Unsupervised Image-to-Image Translation (2018, background): “From another perspective, we note that L2 recon,L recon,L x1 GAN coincide with the objective of a WAE [3] or AAE [4] in the latent space, which pushes the encoded latent distribution towards the latent distribution at generation time.”
Generative Image Modeling Using Style and Structure Adversarial Networks (2016, background): “[43] proposed an adversarial auto-encoder which takes the adversarial loss as an extra constraint for the latent code during training the auto-encoder.”
Neural Variational Inference for Text Processing (2015, background): “The recent variants of generative auto-encoder (Louizos et al., 2015; Makhzani et al., 2015) are also very competitive.”
Invertible Conditional GANs for image editing (2016, background): “Additionally, Makhzani et al. [14] and Larsen et al. [15] proposed a similar idea to this paper by combining a VAE and a GAN with promising results.”
DISCO Nets : DISsimilarity COefficients Networks (2016, background): “[15], where the authors propose to regularise autoencoders with an adversarial network.”