Learning Structured Output Representation using Deep Conditional Generative Models

By Kihyuk Sohn, Honglak Lee, Xinchen Yan. 2015

3,936 citations recorded by the source corpus.

How researchers see this paper

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

In Generative Models

Citations in the corpus

High-Resolution Image Synthesis with Latent Diffusion Models (2021, background): “Similar to other types of generative models [55, 80], diffusion models are in principle capable of modeling conditional distributions of the form p(z|y).”
Tutorial on Variational Autoencoders (2016, background): “While a CVAE works quite well for this purpose, unfortunately a regressor actually works quite well also, producing relatively crisp samples.”
Recent advances in convolutional neural networks (2015, background): “ually p(z) is the standard normal distribution [84], discrete distribution [86], or some distributions with geometric interpretation [93]. Following the original VAE, many variants have been proposed [85, 86, 94]. Conditional VAE [86, 94] generates samples from the conditional distribution with x~ ”
Algorithmic recourse under imperfect causal knowledge: a probabilistic approach (2020, methodology): “It is interesting to note, however, that the CVAE - SCM from (13) can be understood as a non-linear extension of the linear Gaussian model with equal error variances considered by [29], for which identiﬁability has been shown.”
Off-Policy Deep Reinforcement Learning without Exploration (2018, methodology): “The VAE is trained on a reconstruction loss, and a KL-divergence term according to the distribution of the latent vectors.”
Physics-Constrained Deep Learning for High-dimensional Surrogate Modeling and Uncertainty Quantification without Labeled Data (2019, neutral): “Instead of learning this conditional density with labeled data [51, 52, 53], we distill it from a reference density pβ(y|x).”
Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems (2019, background): “where KL q φ ( z | x , y ) || p γ ( z | x ) denotes the Kullback– Leibler divergence between the approximate posterior q φ ( z | x , y ) and the prior over the latent variables p γ ( z | x ) parametrized by γ [22,23].”
Stock Price Prediction by Deep Neural Generative Model of News Articles (2018, background): “The DGM is an implementation of generative model on deep neural networks [15]–[17].”
Graphite: Iterative Generative Modeling of Graphs (2018, background): “In recent years, representations learned by latent variable models parameterized by deep neural networks have shown impressive performance on many tasks such as semi-supervised learning and structured prediction [19, 43].”
Incremental Learning from Scratch for Task-Oriented Dialogue Systems (2019, background): “Fortunately, we can obtain its lower bound (Hoffman et al., 2013; Miao et al., 2016; Sohn et al., 2015) as follows:”
Continual Learning in Generative Adversarial Nets (2017, background): “, sampling, and reconstruction/denoising. These models are also amenable to conditional training, where a conditional input (e.g., class label, data from another modality) guides the sampling process [13, 18]. At test time, manually selected conditional inputs lead to samples from the corresponding ”
Lightweight Probabilistic Deep Networks (2018, methodology): “A more recent approach to latent, generative modeling is the reparametrization trick used in variational autoencoders [32], or conditional autoencoders [30, 46].”
Lifelong Generative Modeling (2017, background): “super-resolution [8], structured prediction [9], clustering [10], and pre-training [11].”
Learning Disentangled Representations with Semi-Supervised Deep Generative Models (2017, background): “In this respect our work extends previous efforts to introduce supervision into variational autoencoders [18, 24, 32].”
Variational Inference for Monte Carlo Objectives (2016, methodology): “We will now show that this approach does not bias the resulting estimator.”
Toward Multimodal Image-to-Image Translation (2017, background): “This approach has been popularized in the unconditional setting by VAEs [23] and VAE-GANs [26].”
Learning Latent Plans from Play (2019, background): “[56], autoencoding random experiences extracted from play memory through a latent plan space.”