ExGAN: Adversarial Generation of Extreme Samples

Abstract

Mitigating the risk arising from extreme events is a fundamental goal with many applications, such as the modelling of natural disasters, financial crashes, epidemics, and many others. To manage this risk, a vital step is to be able to understand or generate a wide range of extreme scenarios. Existing approaches based on Generative Adversarial Networks (GANs) excel at generating realistic samples, but seek to generate typical samples, rather than extreme samples. Hence, in this work, we propose ExGAN, a GAN-based approach to generate realistic and extreme samples. To model the extremes of the training distribution in a principled way, our work draws from Extreme Value Theory (EVT), a probabilistic approach for modelling the extreme tails of distributions. For practical utility, our framework allows the user to specify both the desired extremeness measure, as well as the desired extremeness probability they wish to sample at. Experiments on real US Precipitation data show that our method generates realistic samples, based on visual inspection and quantitative measures, in an efficient manner. Moreover, generating increasingly extreme examples using ExGAN can be done in constant time (with respect to the extremeness probability τ), as opposed to the O(1/τ) time required by the baseline approach.

Extreme scenario generation for renewable energies

“…For example, ref. [6] proposed a new GAN for generating extreme samples in the context of images by distribution shift. Similarly, ref.…”

“…Inspired by ref. [6], our proposed model consists of three core stages, which are extreme data augmentation, extreme conditions generation and conditional scenario generation. The former two aim to prepare the extreme datasets for the final model fitting process while the last one helps to generate extreme scenarios controlled by given electricity generation conditions.…”

“…Therefore, to cope with this challenge, we proposed a three‐staged approach to generate the extreme scenarios for renewable energies. Specifically, we proposed to utilise the concept of distribution shift method [6] to augment the extreme data. Besides, to fit in the characteristics of the extreme data, we adopt Extreme Value Theory (EVT) to model the tail of the historical data.…”

Extreme heat wave sampling and prediction with analog Markov chain and comparisons with deep learning

“…Consequently, for our problem of heat wave prediction, we will consider linear and nonlinear dimensionality reduction techniques and evaluate the performance of SWG in real versus latent space. This approach is partially motivated by the emergence of generative modeling for climate and weather applications, for example, studies combining deep learning architectures with extreme value theory (EVT) for generating extremes (Bhatia et al, 2021; Boulaguiem et al, 2022) and realistic climate situations (Besombes et al, 2021).…”

Modeling and simulating spatial extremes by combining extreme value theory with generative adversarial networks

“…Huster et al (2021) propose to use heavy-tailed input to overcome this issue. Bhatia et al (2021) propose the ExGAN algorithm that uses conditional GANs to perform importance sampling of extreme scenarios. The main difference to our approach is that ExGAN simulates single extreme events, while our model will rely on block maxima and therefore does not suffer from serial correlation and seasonality issues.…”