While stochastic variational inference is relatively well known for scaling inference in Bayesian probabilistic models, related methods also offer ways to circumnavigate the approximation of analytically intractable expectations. The key challenge in either setting is controlling the variance of gradient estimates: recent work has shown that for continuous latent variables, particularly multivariate Gaussians, this can be achieved by using the gradient of the log posterior. In this paper we apply the same idea to gamma distributed latent variables given gamma variational distributions, enabling straightforward "black box" variational inference in models where sparsity and non-negativity are appropriate. We demonstrate the method on a recently proposed gamma process model for network data, as well as a novel sparse factor analysis. We outperform generic sampling algorithms and the approach of using Gaussian variational distributions on transformed variables.

Approximate Bayesian computation (ABC) methods provide an elaborate approach to Bayesian inference on complex models, including model choice. Both theoretical arguments and simulation experiments indicate, however, that model posterior probabilities may be poorly evaluated by standard ABC techniques. We propose a novel approach based on a machine learning tool named random forests to conduct selection among the highly complex models covered by ABC algorithms. We thus modify the way Bayesian model selection is both understood and operated, in that we rephrase the inferential goal as a classification problem, first predicting the model that best fits the data with random forests and postponing the approximation of the posterior probability of the predicted MAP for a second stage also relying on random forests. Compared with earlier implementations of ABC model choice, the ABC random forest approach offers several potential improvements: (i) it often has a larger discriminative power among the competing models, (ii) it is more robust against the number and choice of statistics summarizing the data, (iii) the computing effort is drastically reduced (with a gain in computation efficiency of at least fifty), and (iv) it includes an approximation of the posterior probability of the selected model. The call to random forests will undoubtedly extend the range of size of datasets and complexity of models that ABC can handle. We illustrate the power of this novel methodology by analyzing controlled experiments as well as genuine population genetics datasets. The proposed methodologies are implemented in the R package abcrf available on the CRAN.

The speed with which a learning algorithm converges as it is presented with more data is a central problem in machine learning --- a fast rate of convergence means less data is needed for the same level of performance. The pursuit of fast rates in online and statistical learning has led to the discovery of many conditions in learning theory under which fast learning is possible. We show that most of these conditions are special cases of a single, unifying condition, that comes in two forms: the central condition for 'proper' learning algorithms that always output a hypothesis in the given model, and stochastic mixability for online algorithms that may make predictions outside of the model. We show that under surprisingly weak assumptions both conditions are, in a certain sense, equivalent. The central condition has a re-interpretation in terms of convexity of a set of pseudoprobabilities, linking it to density estimation under misspecification. For bounded losses, we show how the central condition enables a direct proof of fast rates and we prove its equivalence to the Bernstein condition, itself a generalization of the Tsybakov margin condition, both of which have played a central role in obtaining fast rates in statistical learning. Yet, while the Bernstein condition is two-sided, the central condition is one-sided, making it more suitable to deal with unbounded losses. In its stochastic mixability form, our condition generalizes both a stochastic exp-concavity condition identified by Juditsky, Rigollet and Tsybakov and Vovk's notion of mixability. Our unifying conditions thus provide a substantial step towards a characterization of fast rates in statistical learning, similar to how classical mixability characterizes constant regret in the sequential prediction with expert advice setting.

Observations depending on sums of random variables are common throughout many fields; however, no efficient solution is currently known for performing max-product inference on these sums of general discrete distributions (max-product inference can be used to obtain maximum a posteriori estimates). The limiting step to max-product inference is the max-convolution problem (sometimes presented in log-transformed form and denoted as "infimal convolution", "min-convolution", or "convolution on the tropical semiring"), for which no O(k log(k)) method is currently known. Here I present a O(k log(k)) numerical method for estimating the max-convolution of two nonnegative vectors (e.g., two probability mass functions), where k is the length of the larger vector. This numerical max-convolution method is then demonstrated by performing fast max-product inference on a convolution tree, a data structure for performing fast inference given information on the sum of n discrete random variables in O(n k log(n k) log(n) ) steps (where each random variable has an arbitrary prior distribution on k contiguous possible states). The numerical max-convolution method can be applied to specialized classes of hidden Markov models to reduce the runtime of computing the Viterbi path from n k^2 to n k log(k), and has potential application to the all-pairs shortest paths problem.

With an eye towards human-centered automation, we contribute to the development of a systematic means to infer features of human decision-making from behavioral data. Motivated by the common use of softmax selection in models of human decision-making, we study the maximum likelihood parameter estimation problem for softmax decision-making models with linear objective functions. We present conditions under which the likelihood function is convex. These allow us to provide sufficient conditions for convergence of the resulting maximum likelihood estimator and to construct its asymptotic distribution. In the case of models with nonlinear objective functions, we show how the estimator can be applied by linearizing about a nominal parameter value. We apply the estimator to fit the stochastic UCL (Upper Credible Limit) model of human decision-making to human subject data. We show statistically significant differences in behavior across related, but distinct, tasks.

Deep learning methods have shown great promise in many practical applications, ranging from speech recognition, visual object recognition, to text processing. However, most of the current deep learning methods suffer from scalability problems for large-scale applications, forcing researchers or users to focus on small-scale problems with fewer parameters. In this paper, we consider a well-known machine learning model, deep belief networks (DBNs) that have yielded impressive classification performance on a large number of benchmark machine learning tasks. To scale up DBN, we propose an approach that can use the computing clusters in a distributed environment to train large models, while the dense matrix computations within a single machine are sped up using graphics processors (GPU). When training a DBN, each machine randomly drops out a portion of neurons in each hidden layer, for each training case, making the remaining neurons only learn to detect features that are generally helpful for producing the correct answer. Within our approach, we have developed four methods to combine outcomes from each machine to form a unified model. Our preliminary experiment on the mnst handwritten digit database demonstrates that our approach outperforms the state of the art test error rate.

We present two new statistical machine learning methods designed to learn on fully homomorphic encrypted (FHE) data. The introduction of FHE schemes following Gentry (2009) opens up the prospect of privacy preserving statistical machine learning analysis and modelling of encrypted data without compromising security constraints. We propose tailored algorithms for applying extremely random forests, involving a new cryptographic stochastic fraction estimator, and na\"{i}ve Bayes, involving a semi-parametric model for the class decision boundary, and show how they can be used to learn and predict from encrypted data. We demonstrate that these techniques perform competitively on a variety of classification data sets and provide detailed information about the computational practicalities of these and other FHE methods.

Dirichlet Process(DP) is a Bayesian non-parametric prior for infinite mixture modeling, where the number of mixture components grows with the number of data items. The Hierarchical Dirichlet Process (HDP), is an extension of DP for grouped data, often used for non-parametric topic modeling, where each group is a mixture over shared mixture densities. The Nested Dirichlet Process (nDP), on the other hand, is an extension of the DP for learning group level distributions from data, simultaneously clustering the groups. It allows group level distributions to be shared across groups in a non-parametric setting, leading to a non-parametric mixture of mixtures. The nCRF extends the nDP for multilevel non-parametric mixture modeling, enabling modeling topic hierarchies. However, the nDP and nCRF do not allow sharing of distributions as required in many applications, motivating the need for multi-level non-parametric admixture modeling. We address this gap by proposing multi-level nested HDPs (nHDP) where the base distribution of the HDP is itself a HDP at each level thereby leading to admixtures of admixtures at each level. Because of couplings between various HDP levels, scaling up is naturally a challenge during inference. We propose a multi-level nested Chinese Restaurant Franchise (nCRF) representation for the nested HDP, with which we outline an inference algorithm based on Gibbs Sampling. We evaluate our model with the two level nHDP for non-parametric entity topic modeling where an inner HDP creates a countably infinite set of topic mixtures and associates them with author entities, while an outer HDP associates documents with these author entities. In our experiments on two real world research corpora, the nHDP is able to generalize significantly better than existing models and detect missing author entities with a reasonable level of accuracy.

We investigate a Gaussian mixture model (GMM) with component means constrained in a pre-selected subspace. Applications to classification and clustering are explored. An EM-type estimation algorithm is derived. We prove that the subspace containing the component means of a GMM with a common covariance matrix also contains the modes of the density and the class means. This motivates us to find a subspace by applying weighted principal component analysis to the modes of a kernel density and the class means. To circumvent the difficulty of deciding the kernel bandwidth, we acquire multiple subspaces from the kernel densities based on a sequence of bandwidths. The GMM constrained by each subspace is estimated; and the model yielding the maximum likelihood is chosen. A dimension reduction property is proved in the sense of being informative for classification or clustering. Experiments on real and simulated data sets are conducted to examine several ways of determining the subspace and to compare with the reduced rank mixture discriminant analysis (MDA). Our new method with the simple technique of spanning the subspace only by class means often outperforms the reduced rank MDA when the subspace dimension is very low, making it particularly appealing for visualization.

A recurring problem when building probabilistic latent variable models is regularization and model selection, for instance, the choice of the dimensionality of the latent space. In the context of belief networks with latent variables, this problem has been adressed with Automatic Relevance Determination (ARD) employing Monte Carlo inference. We present a variational inference approach to ARD for Deep Generative Models using doubly stochastic variational inference to provide fast and scalable learning. We show empirical results on a standard dataset illustrating the effects of contracting the latent space automatically. We show that the resulting latent representations are significantly more compact without loss of expressive power of the learned models.