Sampling from the posterior distribution using Markov chain Monte Carlo (MCMC) methods can require an exhaustive number of iterations to fully explore the correct posterior. This is often the case when the posterior of interest is multi-modal, as the MCMC sampler can become trapped in a local mode for a large number of iterations. In this paper, we introduce the pseudo-extended MCMC method as an approach for improving the mixing of the MCMC sampler in complex posterior distributions. The pseudo-extended method augments the state-space of the posterior using pseudo-samples as auxiliary variables, where on the extended space, the MCMC sampler is able to easily move between the well-separated modes of the posterior. We apply the pseudo-extended method within an Hamiltonian Monte Carlo sampler and show that by using the No U-turn algorithm (Hoffman and Gelman, 2014), our proposed sampler is completely tuning free. We compare the pseudo-extended method against well-known tempered MCMC algorithms and show the advantages of the new sampler on a number of challenging examples from the statistics literature.

Natural-language-facilitated human-robot cooperation (NLC) refers to using natural language (NL) to facilitate interactive information sharing and task executions with a common goal constraint between robots and humans. Recently, NLC research has received increasing attention. Typical NLC scenarios include robotic daily assistance, robotic health caregiving, intelligent manufacturing, autonomous navigation, and robot social accompany. However, a thorough review, that can reveal latest methodologies to use NL to facilitate human-robot cooperation, is missing. In this review, a comprehensive summary about methodologies for NLC is presented. NLC research includes three main research focuses: NL instruction understanding, NL-based execution plan generation, and knowledge-world mapping. In-depth analyses on theoretical methods, applications, and model advantages and disadvantages are made. Based on our paper review and perspective, potential research directions of NLC are summarized.

Since computers can automate such processes, automatic music generation has become a small, but steadily emerging field in Artificial Intelligence and Machine Learning. Nevertheless, automatic music generation as a problem is far from solved: musical outputs created by artificial systems are regarded as a curiosity by human listeners at best, but all too often they are taken as a direct offense to our sense of musical aesthetics. This sensitivity to violations of even the most subtle musical norms illustrates how complex the problem of (especially polyphonic) music generation is. In addition, there are hardly any objective evaluation criteria to rigorously test and compare music generation systems. This is lamentable, not least since successful methods for automatic music generation would be of considerable commercial interest to the music, gaming, and film industries.

Microsoft is working on an autonomous aircraft that can fly for long periods of time without any power. Rather than a motor, the glider relies on artificial intelligence that mimics how birds fly, autonomously finding thermals, or invisible columns of air that rise due to heat, to carry it for long distances. A recent test of the 16.5-foot aircraft - called a sailplane - conducted in the middle of the desert in Hawthorne, Nevada proved successful - the algorithms researchers developed to help it predict where thermals would appear next worked and kept it aloft. Microsoft is working on an autonomous aircraft that can fly for long periods of time without any power. 'Birds do this seamlessly, and all they're doing is harnessing nature,' Ashish Kapoor, a principal researcher at Microsoft, said.

A central problem in statistics is model selection, the choice between competing models of a stochastic process whose observables are corrupted by noise. In the information-based paradigm of inference, model selection is performed by estimating the predictive performance of the com- peting models. The candidate model with the best estimated predictive performance is selected. Information-based inference is dependent on the accuracy of the estimate of the predictive complexity, a measure of the flexibility of the model in fitting the data. A large-sample-size approximation for the performance is the Akaike Information Criterion (AIC). The AIC approximation fails in a wide range of important applications, either significantly under or over-estimating the complexity. We introduce an improved approximation for the complexity which we use to define a new information criterion: the frequentist information criterion (FIC). FIC extends the applicability of information-based infer- ence to the finite-sample-size regime of regular models and to singular models. We demonstrate the power of the approach in a number of example problems.

This survey explores Procedural Content Generation via Machine Learning (PCGML), defined as the generation of game content using machine learning models trained on existing content. As the importance of PCG for game development increases, researchers explore new avenues for generating high-quality content with or without human involvement; this paper addresses the relatively new paradigm of using machine learning (in contrast with search-based, solver-based, and constructive methods). We focus on what is most often considered functional game content such as platformer levels, game maps, interactive fiction stories, and cards in collectible card games, as opposed to cosmetic content such as sprites and sound effects. In addition to using PCG for autonomous generation, co-creativity, mixed-initiative design, and compression, PCGML is suited for repair, critique, and content analysis because of its focus on modeling existing content. We discuss various data sources and representations that affect the resulting generated content. Multiple PCGML methods are covered, including neural networks, long short-term memory (LSTM) networks, autoencoders, and deep convolutional networks; Markov models, $n$-grams, and multi-dimensional Markov chains; clustering; and matrix factorization. Finally, we discuss open problems in the application of PCGML, including learning from small datasets, lack of training data, multi-layered learning, style-transfer, parameter tuning, and PCG as a game mechanic.

Missing data and noisy observations pose significant challenges for reliably predicting events from irregularly sampled multivariate time series (longitudinal) data. Imputation methods, which are typically used for completing the data prior to event prediction, lack a principled mechanism to account for the uncertainty due to missingness. Alternatively, state-of-the-art joint modeling techniques can be used for jointly modeling the longitudinal and event data and compute event probabilities conditioned on the longitudinal observations. These approaches, however, make strong parametric assumptions and do not easily scale to multivariate signals with many observations. Our proposed approach consists of several key innovations. First, we develop a flexible and scalable joint model based upon sparse multiple-output Gaussian processes. Unlike state-of-the-art joint models, the proposed model can explain highly challenging structure including non-Gaussian noise while scaling to large data. Second, we derive an optimal policy for predicting events using the distribution of the event occurrence estimated by the joint model. The derived policy trades-off the cost of a delayed detection versus incorrect assessments and abstains from making decisions when the estimated event probability does not satisfy the derived confidence criteria. Experiments on a large dataset show that the proposed framework significantly outperforms state-of-the-art techniques in event prediction.

It is well established that neural networks with deep architectures perform better than shallow networks for many tasks in machine learning. In statistical physics, while there has been recent interest in representing physical data with generative modelling, the focus has been on shallow neural networks. A natural question to ask is whether deep neural networks hold any advantage over shallow networks in representing such data. We investigate this question by using unsupervised, generative graphical models to learn the probability distribution of a two-dimensional Ising system. Deep Boltzmann machines, deep belief networks, and deep restricted Boltzmann networks are trained on thermal spin configurations from this system, and compared to the shallow architecture of the restricted Boltzmann machine. We benchmark the models, focussing on the accuracy of generating energetic observables near the phase transition, where these quantities are most difficult to approximate. Interestingly, after training the generative networks, we observe that the accuracy essentially depends only on the number of neurons in the first hidden layer of the network, and not on other model details such as network depth or model type. This is evidence that shallow networks are more efficient than deep networks at representing physical probability distributions associated with Ising systems near criticality.

Deep learning techniques have been hugely successful for traditional supervised and unsupervised machine learning problems. In large part, these techniques solve continuous optimization problems. Recently however, discrete generative deep learning models have been successfully used to efficiently search high-dimensional discrete spaces. These methods work by representing discrete objects as sequences, for which powerful sequence-based deep models can be employed. Unfortunately, these techniques are significantly hindered by the fact that these generative models often produce invalid sequences. As a step towards solving this problem, we propose to learn a deep recurrent validator model. Given a partial sequence, our model learns the probability of that sequence occurring as the beginning of a full valid sequence. Thus this identifies valid versus invalid sequences and crucially it also provides insight about how individual sequence elements influence the validity of discrete objects. To learn this model we propose an approach inspired by seminal work in Bayesian active learning. On a synthetic dataset, we demonstrate the ability of our model to distinguish valid and invalid sequences. We believe this is a key step toward learning generative models that faithfully produce valid discrete objects.

Topic models, and more specifically the class of Latent Dirichlet Allocation (LDA), are widely used for probabilistic modeling of text. MCMC sampling from the posterior distribution is typically performed using a collapsed Gibbs sampler. We propose a parallel sparse partially collapsed Gibbs sampler and compare its speed and efficiency to state-of-the-art samplers for topic models on five well-known text corpora of differing sizes and properties. In particular, we propose and compare two different strategies for sampling the parameter block with latent topic indicators. The experiments show that the increase in statistical inefficiency from only partial collapsing is smaller than commonly assumed, and can be more than compensated by the speedup from parallelization and sparsity on larger corpora. We also prove that the partially collapsed samplers scale well with the size of the corpus. The proposed algorithm is fast, efficient, exact, and can be used in more modeling situations than the ordinary collapsed sampler.