Adversarial learning of probabilistic models has recently emerged as a promising alternative to maximum likelihood. Implicit models such as generative adversarial networks (GAN) often generate better samples compared to explicit models trained by maximum likelihood. Yet, GANs sidestep the characterization of an explicit density which makes quantitative evaluations challenging. To bridge this gap, we propose Flow-GANs, a generative adversarial network for which we can perform exact likelihood evaluation, thus supporting both adversarial and maximum likelihood training. When trained adversarially, Flow-GANs generate high-quality samples but attain extremely poor log-likelihood scores, inferior even to a mixture model memorizing the training data; the opposite is true when trained by maximum likelihood. Results on MNIST and CIFAR-10 demonstrate that hybrid training can attain high held-out likelihoods while retaining visual fidelity in the generated samples.

With the development of free, open-source machine learning and artificial intelligence tools like Google's TensorFlow and sci-kit learn, as well as "ML-as-a-service" products like Google's Cloud Prediction API and Microsoft's Azure Machine Learning platform, it's never been easier for companies of all sizes to harness the power of data. But machine learning is such a vast, complex field. Where do you start learning how to use it in your business? In this article, we'll survey the current landscape of machine learning algorithms and explain how they work, provide example applications, share how other companies use them, and provide further resources on learning about them. This executive overview will provide the first step in learning how to apply machine learning algorithm(s) to make your business more efficient, more effective, and more profitable.

Feature extraction for automatic classification of EEG signals typically relies on time frequency representations of the signal. Techniques such as cepstral-based filter banks or wavelets are popular analysis techniques in many signal processing applications including EEG classification. In this paper, we present a comparison of a variety of approaches to estimating and postprocessing features. To further aid in discrimination of periodic signals from aperiodic signals, we add a differential energy term. We evaluate our approaches on the TUH EEG Corpus, which is the largest publicly available EEG corpus and an exceedingly challenging task due to the clinical nature of the data. We demonstrate that a variant of a standard filter bank-based approach, coupled with first and second derivatives, provides a substantial reduction in the overall error rate. The combination of differential energy and derivatives produces a 24% absolute reduction in the error rate and improves our ability to discriminate between signal events and background noise. This relatively simple approach proves to be comparable to other popular feature extraction approaches such as wavelets, but is much more computationally efficient.

Clinical electroencephalographic (EEG) data varies significantly depending on a number of operational conditions (e.g., the type and placement of electrodes, the type of electrical grounding used). This investigation explores the statistical differences present in two different referential montages: Linked Ear (LE) and Averaged Reference (AR). Each of these accounts for approximately 45% of the data in the TUH EEG Corpus. In this study, we explore the impact this variability has on machine learning performance. We compare the statistical properties of features generated using these two montages, and explore the impact of performance on our standard Hidden Markov Model (HMM) based classification system. We show that a system trained on LE data significantly outperforms one trained only on AR data (77.2% vs. 61.4%). We also demonstrate that performance of a system trained on both data sets is somewhat compromised (71.4% vs. 77.2%). A statistical analysis of the data suggests that mean, variance and channel normalization should be considered. However, cepstral mean subtraction failed to produce an improvement in performance, suggesting that the impact of these statistical differences is subtler.

Transfer learning has recently attracted significant research attention, as it simultaneously learns from different source domains, which have plenty of labeled data, and transfers the relevant knowledge to the target domain with limited labeled data to improve the prediction performance. We propose a Bayesian transfer learning framework where the source and target domains are related through the joint prior density of the model parameters. The modeling of joint prior densities enables better understanding of the "transferability" between domains. We define a joint Wishart density for the precision matrices of the Gaussian feature-label distributions in the source and target domains to act like a bridge that transfers the useful information of the source domain to help classification in the target domain by improving the target posteriors. Using several theorems in multivariate statistics, the posteriors and posterior predictive densities are derived in closed forms with hypergeometric functions of matrix argument, leading to our novel closed-form and fast Optimal Bayesian Transfer Learning (OBTL) classifier. Experimental results on both synthetic and real-world benchmark data confirm the superb performance of the OBTL compared to the other state-of-the-art transfer learning and domain adaptation methods.

Predictive modelling and supervised learning are central to modern data science. With predictions from an ever-expanding number of supervised black-box strategies - e.g., kernel methods, random forests, deep learning aka neural networks - being employed as a basis for decision making processes, it is crucial to understand the statistical uncertainty associated with these predictions. As a general means to approach the issue, we present an overarching framework for black-box prediction strategies that not only predict the target but also their own predictions' uncertainty. Moreover, the framework allows for fair assessment and comparison of disparate prediction strategies. For this, we formally consider strategies capable of predicting full distributions from feature variables, so-called probabilistic supervised learning strategies. Our work draws from prior work including Bayesian statistics, information theory, and modern supervised machine learning, and in a novel synthesis leads to (a) new theoretical insights such as a probabilistic bias-variance decomposition and an entropic formulation of prediction, as well as to (b) new algorithms and meta-algorithms, such as composite prediction strategies, probabilistic boosting and bagging, and a probabilistic predictive independence test. Our black-box formulation also leads (c) to a new modular interface view on probabilistic supervised learning and a modelling workflow API design, which we have implemented in the newly released skpro machine learning toolbox, extending the familiar modelling interface and meta-modelling functionality of sklearn. The skpro package provides interfaces for construction, composition, and tuning of probabilistic supervised learning strategies, together with orchestration features for validation and comparison of any such strategy - be it frequentist, Bayesian, or other.

We study the sparse non-negative least squares (S-NNLS) problem. S-NNLS occurs naturally in a wide variety of applications where an unknown, non-negative quantity must be recovered from linear measurements. We present a unified framework for S-NNLS based on a rectified power exponential scale mixture prior on the sparse codes. We show that the proposed framework encompasses a large class of S-NNLS algorithms and provide a computationally efficient inference procedure based on multiplicative update rules. Such update rules are convenient for solving large sets of S-NNLS problems simultaneously, which is required in contexts like sparse non-negative matrix factorization (S-NMF). We provide theoretical justification for the proposed approach by showing that the local minima of the objective function being optimized are sparse and the S-NNLS algorithms presented are guaranteed to converge to a set of stationary points of the objective function. We then extend our framework to S-NMF, showing that our framework leads to many well known S-NMF algorithms under specific choices of prior and providing a guarantee that a popular subclass of the proposed algorithms converges to a set of stationary points of the objective function. Finally, we study the performance of the proposed approaches on synthetic and real-world data.

Statistical performance bounds for reinforcement learning (RL) algorithms can be critical for high-stakes applications like healthcare. This paper introduces a new framework for theoretically measuring the performance of such algorithms called Uniform-PAC, which is a strengthening of the classical Probably Approximately Correct (PAC) framework. In contrast to the PAC framework, the uniform version may be used to derive high probability regret guarantees and so forms a bridge between the two setups that has been missing in the literature. We demonstrate the benefits of the new framework for finite-state episodic MDPs with a new algorithm that is Uniform-PAC and simultaneously achieves optimal regret and PAC guarantees except for a factor of the horizon.

So, what are Markov chain Monte Carlo (MCMC) methods? In this article, I will explain that short answer, without any math. A parameter of interest is just some number that summarizes a phenomenon we're interested in. In general we use statistics to estimate parameters. For example, if we want to learn about the height of human adults, our parameter of interest might be average height in in inches.

In this work, we try to understand the differences between exact and approximate inference algorithms in structured prediction. We compare the estimation and approximation error of both underestimate (e.g., greedy search) and overestimate (e.g., linear relaxation of integer programming) models. The result shows that, from the perspective of learning errors, performances of approximate inference could be as good as exact inference. The error analyses also suggest a new margin for existing learning algorithms. Empirical evaluations on text classification, sequential labelling and dependency parsing witness the success of approximate inference and the benefit of the proposed margin.