Estimation with Uncertainty via Conditional Generative Adversarial Networks

Abstract

Conventional predictive Artificial Neural Networks (ANNs) commonly employ deterministic weight matrices; therefore, their prediction is a point estimate. Such a deterministic nature in ANNs causes the limitations of using ANNs for medical diagnosis, law problems, and portfolio management in which not only discovering the prediction but also the uncertainty of the prediction is essentially required. In order to address such a problem, we propose a predictive probabilistic neural network model, which corresponds to a different manner of using the generator in the conditional Generative Adversarial Network (cGAN) that has been routinely used for conditional sample generation. By reversing the input and output of ordinary cGAN, the model can be successfully used as a predictive model; moreover, the model is robust against noises since adversarial training is employed. In addition, to measure the uncertainty of predictions, we introduce the entropy and relative entropy for regression problems and classification problems, respectively. The proposed framework is applied to stock market data and an image classification task. As a result, the proposed framework shows superior estimation performance, especially on noisy data; moreover, it is demonstrated that the proposed framework can properly estimate the uncertainty of predictions.

An Ensemble Deep Learning Model with a Gene Attention Mechanism for Estimating the Prognosis of Low-Grade Glioma

“…Owing to the rapid development of deep learning algorithms, artificial intelligence models based on deep learning have become mainstream in various domains [1][2][3]. In biomedicine, such models have been extensively studied for medical imaging [4][5][6], diagnosis [7][8][9], and genome sequencing [10][11][12].…”

Estimating the Prognosis of Low-Grade Glioma with Gene Attention Using Multi-Omics and Multi-Modal Schemes

“…Owing to the rapid development of deep learning models, artificial intelligence based on deep learning has dominated various fields [ 1 , 2 , 3 ]. Deep learning models are being used extensively in diagnostics [ 4 , 5 , 6 ], medical imaging [ 7 , 8 , 9 , 10 ], and genome sequencing [ 11 , 12 , 13 ].…”

Probabilistic Wind Park Power Prediction using Bayesian Deep Learning and Generative Adversarial Networks

“…Generative Adversarial Networks (GAN) were first introduced by Goodfellow et al [24] and have since become the quintessential DL architecture for image generation. Recently, GANs have also been proposed as probabilistic prediction models [25,26]. Adler et al [25] used a GAN for image reconstruction of computed tomography images.…”

“…By sampling from the model, the authors managed to obtain probabilistic predictions, with point-wise mean and standard deviation estimates for the reconstructed images. Similarly, Lee et al [26] used a conditional GAN as a probabilistic regression model to estimate the uncertainty, or risk, associated with an expected return prediction for the stock market. The main contributions of this paper can be summarised as follows:…”