Adversarially Learned Inference

By Vincent Dumoulin, Ishmael Belghazi, Ben Poole, Alex Lamb, Martín Arjovsky, Olivier Mastropietro, Aaron C. Courville. 2016

1,346 citations recorded by the source corpus.

We introduce the adversarially learned inference (ALI) model, which jointly learns a generation network and an inference network using an adversarial process. The generation network maps samples from stochastic latent variables to the data space while the inference network maps training examples in data space to the space of latent variables. An adversarial game is cast between these two networks and a discriminative network is trained to distinguish between joint latent/data-space samples from the generative network and joint samples from the inference network. We illustrate the ability of the model to learn mutually coherent inference and generation networks through the inspections of model samples and reconstructions and confirm the usefulness of the learned representations by obtaining a performance competitive with state-of-the-art on the semi-supervised SVHN and CIFAR10 tasks.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Statistical and Nonlinear Physics, Generative Models, Probabilistic & Bayesian Methods

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Statistical and Nonlinear Physics

In Generative Models

In Probabilistic & Bayesian Methods

Citations in the corpus

Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning (2020, background): “Many of these approaches rely either on auto-encoding of images [24, 25, 26] or on adversarial learning [27], jointly modelling data and representation [28, 29, 30, 31].”
Learning Features by Watching Objects Move (2016, background): “These models can be extended to produce good feature representations by training jointly with image encoders [10,11].”
A Style-Based Generator Architecture for Generative Adversarial Networks (2018, background): “While it is possible to add an extra network for this purpose [7, 11, 14], we want to avoid investing effort into a component that is not a part of the actual solution.”
Autoregressive Quantile Networks for Generative Modeling (2018, background): “Many recent advances have come from principled combinations of these three fundamental methods (Makhzani et al., 2015; Dumoulin et al., 2016; Rosca et al., 2017).”
Progressive Growing of GANs for Improved Quality, Stability, and Variation (2017, methodology): “017; Berthelot et al., 2017; Kodali et al., 2017). Hybrid methods combine various strengths of the three, but so far lag behind GANs in image quality (Makhzani & Frey, 2017; Ulyanov et al., 2017; Dumoulin et al., 2016). Typically, a GAN consists of two networks: generator and discriminator (aka ”
D2C: Diffusion-Decoding Models for Few-Shot Conditional Generation (2021, neutral): “Model family Explicit x→ z No prior hole Non-Adversarial (Enables SSL) (Better generation) (Stable training) VAE [52, 77], NF [28] 4 7 4 GAN [34] 7 4 7 BiGAN [30, 32] 4 4 7 DDIM [84] 7 4 4 D2C 4 4 4”
Dual Contradistinctive Generative Autoencoder (2020, neutral): “The subsequent development after GAN [55, 2, 17, 34, 15, 14, 13] has led to a great leap forward in building decoder-based generative models.”
Learning Layout and Style Reconfigurable GANs for Controllable Image Synthesis (2020, methodology): “In [16], [17], [41], conditional information is utilized by the discriminator using simple concatenation with the input or intermediate feature maps.”
Improved Training of Wasserstein GANs (2017, neutral): “Method Score ALI [8] (in [26]) 5 .”
Unrolled Generative Adversarial Networks (2016, methodology): “Once converged, the generative models produced by the GAN training procedure normally do not cover the whole distribution (Dumoulin et al., 2016; Che et al., 2016), even when targeting a mode-covering divergence such as KL. Additionally, because it is intractable to compute the GAN training loss,…”
GAN Inversion: A Survey (2021, methodology): “Although some methods [31], [32], [33], [96] use additive”
Large Scale Adversarial Representation Learning (2019, methodology): “Under this objective, [5, 8] showed that with an optimal D, G and E minimize the Jensen-Shannon divergence between the joint distributions PxE and PGz, and therefore at the global optimum, the two joint distributions PxE = PGz match, analogous to the results from standard GANs [11].”
A critical analysis of self-supervision, or what we can learn from a single image (2019, background): “Bidirectional Generative Adversarial Networks (BiGAN) (Donahue et al., 2017; Dumoulin et al., 2016) are an extension of GANs designed to learn a useful image representation as an approximate inverse of the generator through joint inference on an encoding and the image.”
Unsupervised Embedding Learning via Invariant and Spreading Instance Feature (2019, background): “Bolztmann Machines (RBMs) [24, 40], Auto-encoders [20, 42] and generative adversarial network (GAN) [7, 10, 11] are widely studied.”
Learning deep representations by mutual information estimation and maximization (2018, background): “Overall, all models showed much lower dependence than BiGAN, indicating the marginal of the encoder output is not matching to the generator’s spherical Gaussian input prior.”
Split-Brain Autoencoders: Unsupervised Learning by Cross-Channel Prediction (2016, methodology): “Many of these methods [15, 36, 5, 6, 32] have been evaluated for representation learning.”
The Numerics of GANs (2017, background): “Since their ﬁrst appearance, GANs have been successfully applied to a variety of tasks, including image-to-image translation [10], image super-resolution [11], image in-painting [23] domain adaptation [22], probabilistic inference [12, 7, 6] and many more.”
Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks (2017, methodology): “BiGAN/ALI [8, 6] Unconditional GANs [15] learn a generator G : Z → X , that maps random noise Z to images X .”