We consider the task of spatio-temporal landcover monitoring and formulate it as follows: Given a time sequence of multispectral satellite images of an area of interest (AOI), we want to predict a corresponding sequence of semantic segmentations with segment labels representing landcover types. We propose to combine asymmetric UNets (achieving super-resolution segmentation) with Markov chain models to account for both spatial and temporal dependencies. Such models cannot be trained in a supervised manner, as obtaining dense spatio-temporal annotations for satellite image time sequences is infeasible. We therefore focus on the challenge of their semi-supervised training. The proposed approach is evaluated on the task of forest monitoring in a national park in the Czech Republic, which suffers a severe forest dieback due to droughts and bark beetle outbreaks. We achieve 83% (90%) prediction accuracy in this challenging task.

We view variational autoencoders (VAE) as decoder-encoder pairs, which map distributions in the data space to distributions in the latent space and vice versa. The standard learning approach for VAEs is the maximisation of the evidence lower bound (ELBO). It is asymmetric in that it aims at learning a latent variable model while using the encoder as an auxiliary means only. Moreover, it requires a closed form a-priori latent distribution. This limits its applicability in more complex scenarios, such as general semi-supervised learning and employing complex generative models as priors. We propose a Nash equilibrium learning approach, which is symmetric with respect to the encoder and decoder and allows learning VAEs in situations where both the data and the latent distributions are accessible only by sampling. The flexibility and simplicity of this approach allows its application to a wide range of learning scenarios and downstream tasks.

In neural networks with binary activations and or binary weights the training by gradient descent is complicated as the model has piecewise constant response. We consider stochastic binary networks, obtained by adding noises in front of activations. The expected model response becomes a smooth function of parameters, its gradient is well defined but it is challenging to estimate it accurately. We propose a new method for this estimation problem combining sampling and analytic approximation steps. The method has a significantly reduced variance at the price of a small bias which gives a very practical tradeoff in comparison with existing unbiased and biased estimators. We further show that one extra linearization step leads to a deep straight-through estimator previously known only as an ad-hoc heuristic. We experimentally show higher accuracy in gradient estimation and demonstrate a more stable and better performing training in deep convolutional models with both proposed methods.

The aim of this short note is to draw attention to a method by which the partition function and marginal probabilities for a certain class of random fields on complete graphs can be computed in polynomial time. This class includes Ising models with homogeneous pairwise potentials but arbitrary (inhomogeneous) unary potentials. Similarly, the partition function and marginal probabilities can be computed in polynomial time for random fields on complete bipartite graphs, provided they have homogeneous pairwise potentials. We expect that these tractable classes of large-scale random fields can be very useful for the evaluation of approximation algorithms by providing exact error estimates.

We analyse the potential of Gibbs Random Fields for shape prior modelling. We show that the expressive power of second order GRFs is already sufficient to express spatial relations between shape parts and simple shapes simultaneously. This allows to model and recognise complex shapes as spatial compositions of simpler parts.