Generative adversarial networks
By Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron C. Courville, Yoshua Bengio. Communications of the ACM, 2014
4,455 citations recorded by the source corpus.
Generative adversarial networks are a kind of artificial intelligence algorithm designed to solve the generative modeling problem. The goal of a generative model is to study a collection of training examples and learn the probability distribution that generated them. Generative Adversarial Networks (GANs) are then able to generate more examples from the estimated probability distribution. Generative models based on deep learning are common, but GANs are among the most successful generative models (especially in terms of their ability to generate realistic high-resolution images). GANs have been successfully applied to a wide variety of tasks (mostly in research settings) but continue to present unique challenges and research opportunities because they are based on game theory while most other approaches to generative modeling are based on optimization.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Generative Models
View Paper: https://arxiv.org/abs/1406.2661
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
- Jun-Yan Zhu
- Felix Stahlberg
- Michael Arbel
- Jan Kautz
- Arthur Gretton
- Dacheng Tao
- Fan Yang
- Danica J. Sutherland
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Ming-Hsuan Yang
- Trevor Darrell
- Xiaogang Wang
- Jitendra Malik
- Antonio Torralba
- Bernt Schiele
In Generative Models
- Jun-Yan Zhu
- Ming-Yu Liu
- Jan Kautz
- Ting-Chun Wang
- Andrew Tao
- Bryan Catanzaro
- Richard Zhang
- Eli Shechtman
Citations in the corpus
- Neural Tangent Kernel: Convergence and Generalization in Neural Networks (2018, neutral): “ p F (L)((t));d t E pin : In the case of gradient descent, d t = dj f (t) (see Section 3), but the direction may depend on another network, as is the case for e.g. Generative Adversarial Networks [9]. We only assume that the integral R T 0 kd tk pindtstays stochastically bounded as the width tends ”
- High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs (2017, positive): “nce of existing objects. To synthesize images from semantic labels, one can use the pix2pix method, an image-to-image translation framework [21] which leverages generative adversarial networks (GANs) [16] in a conditional setting. Recently, Chen and Koltun [5] suggest that adversarial training might”
- Demystifying MMD GANs (2018, positive): “Generative Adversarial Networks (GANs; Goodfellow et al., 2014) provide a powerful method for general-purpose generative modeling of datasets.”
- Faster AutoAugment: Learning Augmentation Strategies using Backpropagation (2019, neutral): “nimize error rate, or equivalently to maximize accuracy, while [23, 18] try to match the densities of augmented and original images. Another direction is to use generative adversarial networks (GANs) [9]. [30, 1] use conditional GANs to gen Figure 3. Schematic view of the problem setting. Eac”
- PixelVAE++: Improved PixelVAE with Discrete Prior (2019, positive): “Generative modeling of images [1, 2, 3, 4, 5], audio [6, 7], and videos [8, 9] has advanced remarkably in the past few years and have resulted in great applications [10, 11].”
- Data-centric Artificial Intelligence: A Survey (2023, positive): “GAN [81, 265] has been widely used for data augmentation [74].”
- A Transdisciplinary Review of Deep Learning Research and Its Relevance for Water Resources Scientists (2017, positive): “Generative adversarial networks (GAN) was a progress milestone in DL (Goodfellow et al., 2014; Mogren, 2016) motivated by the need to model high-dimensional, multimodal distributions.”
- Machine learning for molecular and materials science (2018, positive): “80 Molecular science is benefitting from cutting edge algorithmic developments in ML such as generative adversarial networks (GANs) 81 and reinforcement learning for the computational design of novel putative biologically active compounds.”
- Deep learning in biomedicine (2018, neutral): “Generative adversarial networks 22 , a recent innovation in generative modeling, aim to generate examples that are indistinguishable from the real examples the model was trained on.”
- ExpandNet: A Deep Convolutional Neural Network for High Dynamic Range Expansion from Low Dynamic Range Content (2018, positive): “In image-to-image translation [IZZE16b] the authors present a method based on Generative Adversarial Networks [GPAM ∗ 14] and the U-Net [RFB15] architecture that transforms images from one domain to another (e.g. maps to satellite).”
- Convolutional neural network: a review of models, methodologies and applications to object detection (2019, positive): “Another one is generative adversarial networks (GAN) [6] which enable to tackle the problem of unsupervised learning.”
- Retrieval-Augmented Generation for AI-Generated Content: A Survey (2024, neutral): “4) GAN : Generative Adversarial Networks (GANs) [14] are highly anticipated deep learning models with amazing capabilities which can simulate and generate realistic images, audio, and other data.”
- Playing repeated games with large language models (2023, neutral): “Finally, the theory behind interacting agents is important for many machine learning applications in general [Crandall and Goodrich, 2011], and in particular, in adversarial settings [Goodfellow et al., 2020], where one agent tries to trick the other agent into thinking that a generated output is go”