We propose a novel theoretical framework of analysis for Generative Adversarial Networks (GANs).
Both are sub-optimal: the former disregards the discrepancies between environments leading to biased solutions, while the latter does not exploit their potential commonalities and is prone to scarcity problems.
We propose a Monte Carlo objective that leverages the conditional linearity by computing the corresponding conditional expectations in closed-form and a suitable proposal distribution that is factorised similarly to the optimal proposal distribution.
no code implementations • • Emmanuel de Bézenac, Syama Sundar Rangapuram, Konstantinos Benidis, Michael Bohlke-Schneider, Richard Kurle, Lorenzo Stella, Hilaf Hasson, Patrick Gallinari, Tim Januschowski
This paper tackles the modelling of large, complex and multivariate time series panels in a probabilistic setting.
In this work, we introduce the APHYNITY framework, a principled approach for augmenting incomplete physical dynamics described by differential equations with deep data-driven models.
From this observation, we reformulate the learning problem as follows: finding neural networks which solve the task while transporting the data as efficiently as possible.
We tackle the problem of inpainting occluded area in spatiotemporal sequences, such as cloud occluded satellite observations, in an unsupervised manner.
We consider the problem of forecasting complex, nonlinear space-time processes when observations provide only partial information of on the system's state.