This study investigates the effects of Markov Chain Monte Carlo (MCMC) sampling in unsupervised Maximum Likelihood (ML) learning. Our attention is restricted to the family of unnormalized probability densities for which the negative log density (or energy function) is a ConvNet. In general, we find that many of the techniques used to stabilize training in previous studies can have the opposite effect. Stable ML learning with a ConvNet potential can be achieved with only a few hyper-parameters and no regularization. Using this minimal framework, we identify a variety of ML learning outcomes that depend on the implementation of MCMC sampling. On one hand, we show that it is easy to train an energy-based model which can sample realistic images with short-run Langevin. ML can be effective and stable even when MCMC samples have much higher energy than true steady-state samples throughout training. Based on this insight, we introduce an ML method with purely noise-initialized MCMC, high-quality short-run synthesis, and the same budget as ML with informative MCMC initialization such as CD or PCD. Unlike previous models, our model can obtain realistic high-diversity samples from a noise signal after training with no auxiliary networks. On the other hand, ConvNet potentials learned with highly non-convergent MCMC do not have a valid steady-state and cannot be considered approximate unnormalized densities of the training data because long-run MCMC samples differ greatly from observed images. We show that it is much harder to train a ConvNet potential to learn a steady-state over realistic images. To our knowledge, long-run MCMC samples of all previous models lose the realism of short-run samples. With correct tuning of Langevin noise, we train the first ConvNet potentials for which long-run and steady-state MCMC samples are realistic images.

This paper revisits the Bayesian CMA-ES and provides updates for normal Wishart. It emphasizes the difference between a normal and normal inverse Wishart prior. After some computation, we prove that the only difference relies surprisingly in the expected covariance. We prove that the expected covariance should be lower in the normal Wishart prior model because of the convexity of the inverse. We present a mixture model that generalizes both normal Wishart and normal inverse Wishart model. We finally present various numerical experiments to compare both methods as well as the generalized method.

Young isolated neutron stars (INS) most commonly manifest themselves as rotationally powered pulsars (RPPs) which involve conventional radio pulsars as well as gamma-ray pulsars (GRPs) and rotating radio transients (RRATs). Some other young INS families manifest themselves as anomalous X-ray pulsars (AXPs) and soft gamma-ray repeaters (SGRs) which are commonly accepted as magnetars, i.e.\ magnetically powered neutron stars with decaying super-strong fields. Yet some other young INS are identified as central compact objects (CCOs) and X-ray dim isolated neutron stars (XDINs) which are cooling objects powered by their thermal energy. Older pulsars, as a result of a previous long episode of accretion from a companion, manifest themselves as millisecond pulsars and more commonly appear in binary systems. We use Dirichlet process Gaussian mixture model (DPGMM), an unsupervised machine learning algorithm, for analyzing the distribution of these pulsar families in period $P$ and period derivative $\dot{P}$ parameter space. We compare the average values of the characteristic age, magnetic dipole field strength, surface temperature and proper motion of all discovered components. We verify that DPGMM is robust and provides hints for inferring relations between different classes of pulsars. We discuss the implications of our findings for the magnetothermal spin evolution models and fallback discs.

Bayesian inference provides an attractive online-learning framework to analyze sequential data, and offers generalization guarantees which hold even under model mismatch and with adversaries. Unfortunately, exact Bayesian inference is rarely feasible in practice and approximation methods are usually employed, but do such methods preserve the generalization properties of Bayesian inference? In this paper, we show that this is indeed the case for some variational inference (VI) algorithms. We propose new online, tempered VI algorithms and derive their generalization bounds. Our theoretical result relies on the convexity of the variational objective, but we argue that our result should hold more generally and present empirical evidence in support of this. Our work in this paper presents theoretical justifications in favor of online algorithms that rely on approximate Bayesian methods.

Structured latent attribute models (SLAMs) are a special family of discrete latent variable models widely used in social and biological sciences. This paper considers the problem of learning significant attribute patterns from a SLAM with potentially high-dimensional configurations of the latent attributes. We address the theoretical identifiability issue, propose a penalized likelihood method for the selection of the attribute patterns, and further establish the selection consistency in such an overfitted SLAM with diverging number of latent patterns. The good performance of the proposed methodology is illustrated by simulation studies and two real datasets in educational assessment.

The concept of mathematical modeling is widespread across almost all of the fields of contemporary science and engineering. Because of the existing necessity of predictions the behavior of natural phenomena, the researchers develop more and more complex models. However, despite their ability to better forecasting, the problem of an appropriate fitting ground truth data to those, high-dimensional and nonlinear models seems to be inevitable. In order to deal with this demanding problem the entire discipline of data assimilation has been developed. Basing on the Human and Nature Dynamics (HANDY) model, we have presented a detailed and comprehensive comparison of Approximate Bayesian Computation (classic data assimilation method) and a novelty approach of Supermodeling. Furthermore, with the usage of Sensitivity Analysis, we have proposed the methodology to reduce the number of coupling coefficients between submodels and as a consequence to increase the speed of the Supermodel converging. In addition, we have demonstrated that usage of Approximate Bayesian Computation method with the knowledge about parameters' sensitivities could result with satisfactory estimation of the initial parameters. However, we have also presented the mentioned methodology as unable to achieve similar predictions to Approximate Bayesian Computation. Finally, we have proved that Supermodeling with synchronization via the most sensitive variable could effect with the better forecasting for chaotic as well as more stable systems than the Approximate Bayesian Computation. What is more, we have proposed the adequate methodologies.

We recapitulate the Bayesian formulation of neural network based classifiers and show that, while sampling from the posterior does indeed lead to better generalisation than is obtained by standard optimisation of the cost function, even better performance can in general be achieved by sampling finite temperature ($T$) distributions derived from the posterior. Taking the example of two different deep (3 hidden layers) classifiers for MNIST data, we find quite different $T$ values to be appropriate in each case. In particular, for a typical neural network classifier a clear minimum of the test error is observed at $T>0$. This suggests an early stopping criterion for full batch simulated annealing: cool until the average validation error starts to increase, then revert to the parameters with the lowest validation error. As $T$ is increased classifiers transition from accurate classifiers to classifiers that have higher training error than assigning equal probability to each class. Efficient studies of these temperature-induced effects are enabled using a replica-exchange Hamiltonian Monte Carlo simulation technique. Finally, we show how thermodynamic integration can be used to perform model selection for deep neural networks. Similar to the Laplace approximation, this approach assumes that the posterior is dominated by a single mode. Crucially, however, no assumption is made about the shape of that mode and it is not required to precisely compute and invert the Hessian.

In computer vision, image datasets used for classification are naturally associated with multiple labels and comprised of multiple views, because each image may contain several objects (e.g. pedestrian, bicycle and tree) and is properly characterized by multiple visual features (e.g. color, texture and shape). Currently available tools ignore either the label relationship or the view complementary. Motivated by the success of the vector-valued function that constructs matrix-valued kernels to explore the multi-label structure in the output space, we introduce multi-view vector-valued manifold regularization (MV$\mathbf{^3}$MR) to integrate multiple features. MV$\mathbf{^3}$MR exploits the complementary property of different features and discovers the intrinsic local geometry of the compact support shared by different features under the theme of manifold regularization. We conducted extensive experiments on two challenging, but popular datasets, PASCAL VOC' 07 (VOC) and MIR Flickr (MIR), and validated the effectiveness of the proposed MV$\mathbf{^3}$MR for image classification.

This work tackles the problem of learning a set of language specific acoustic units from unlabeled speech recordings given a set of labeled recordings from other languages. Our approach may be described by the following two steps procedure: first the model learns the notion of acoustic units from the labelled data and then the model uses its knowledge to find new acoustic units on the target language. We implement this process with the Bayesian Subspace Hidden Markov Model (SHMM), a model akin to the Subspace Gaussian Mixture Model (SGMM) where each low dimensional embedding represents an acoustic unit rather than just a HMM's state. The subspace is trained on 3 languages from the GlobalPhone corpus (German, Polish and Spanish) and the AUs are discovered on the TIMIT corpus. Results, measured in equivalent Phone Error Rate, show that this approach significantly outperforms previous HMM based acoustic units discovery systems and compares favorably with the Variational Auto Encoder-HMM.

Learning a shared dialog structure from a set of task-oriented dialogs is an important challenge in computational linguistics. The learned dialog structure can shed light on how to analyze human dialogs, and more importantly contribute to the design and evaluation of dialog systems. We propose to extract dialog structures using a modified VRNN model with discrete latent vectors. Different from existing HMM-based models, our model is based on variational-autoencoder (VAE). Such model is able to capture more dynamics in dialogs beyond the surface forms of the language. We find that qualitatively, our method extracts meaningful dialog structure, and quantitatively, outperforms previous models on the ability to predict unseen data. We further evaluate the model's effectiveness in a downstream task, the dialog system building task. Experiments show that, by integrating the learned dialog structure into the reward function design, the model converges faster and to a better outcome in a reinforcement learning setting.