Count-based estimators are a fundamental building block of a number of powerful sequential prediction algorithms, including Context Tree Weighting and Prediction by Partial Matching. Keeping exact counts, however, typically results in a high memory overhead. In particular, when dealing with large alphabets the memory requirements of count-based estimators often become prohibitive. In this paper we propose three novel ideas for approximating count-based estimators using bounded memory. Our first contribution, of independent interest, is an extension of reservoir sampling for sampling distinct symbols from a stream of unknown length, which we call K-distinct reservoir sampling. We combine this sampling scheme with a state-of-the-art count-based estimator for memoryless sources, the Sparse Adaptive Dirichlet (SAD) estimator. The resulting algorithm, the Budget SAD, naturally guarantees a limit on its memory usage. We finally demonstrate the broader use of K-distinct reservoir sampling in nonparametric estimation by using it to restrict the branching factor of the Context Tree Weighting algorithm. We demonstrate the usefulness of our algorithms with empirical results on two sequential, large-alphabet prediction problems.

Trust schemes can identify such nodes, as they Wireless sensor networks are being increasingly can predict a node's behavior (quality) both directly, via evaluation used for sustainable development. The task of routing based on its past actions, and indirectly, using recommendations in these resource-constraint networks is particularly (opinions) from other nodes. However, many challenging as they operate over prolonged trust schemes cannot effectively handle attacks targeting trust deployment periods, necessitating optimal use of systems themselves [Sun et al., 2006] i.e., they are heavily their resources. Moreover, due to the deployment affected by malicious nodes deliberately providing misleading in unattended environments, they become an easy opinions (unfair ratings) about other nodes.

Biclustering is an intrinsically challenging and highly complex problem, particularly studied in the biology field, where the goal is to simultaneously cluster genes and samples of an expression data matrix. In this paper we present a novel approach to gene expression biclustering by providing a binary Factor Graph formulation to such problem. In more detail, we reformulate biclustering as a sequential search for single biclusters and use an efficient optimization procedure based on the Max Sum algorithm. Such approach, drastically alleviates the scaling issues of previous approaches for biclustering based on Factor Graphs obtaining significantly more accurate results on synthetic datasets. A further analysis on two real-world datasets confirms the potentials of the proposed methodology when compared to alternative state of the art methods.

When a transition probability matrix is represented as the product of two stochastic matrices, swapping the factors of the multiplication yields another transition matrix that retains some fundamental characteristics of the original. Since the new matrix can be much smaller than its precursor, replacing the former for the latter can lead to significant savings in terms of computational effort. This strategy, dubbed the "stochastic-factorization trick," can be used to compute the stationary distribution of a Markov chain, to determine the fundamental matrix of an absorbing chain, and to compute a decision policy via dynamic programming or reinforcement learning. In this paper we show that the stochastic-factorization trick can also provide benefits in terms of the number of samples needed to estimate a transition matrix. We introduce a probabilistic interpretation of a stochastic factorization and build on the resulting model to develop an algorithm to compute the factorization directly from data. If the transition matrix can be well approximated by a low-order stochastic factorization, estimating its factors instead of the original matrix reduces significantly the number of parameters to be estimated. Thus, when compared to estimating the transition matrix directly via maximum likelihood, the proposed method is able to compute approximations of roughly the same quality using less data. We illustrate the effectiveness of the proposed algorithm by using it to help a reinforcement learning agent learn how to play the game of blackjack.

Sum-product networks (SPNs) are rooted, directed acyclic graphs (DAGs) of sum and product nodes with well-defined probabilistic semantics. Moreover, exact inference in the distribution represented by an SPN is guaranteed to take linear time in the size of the DAG. In this paper we introduce an algorithm that learns the structure of an SPN using a greedy search approach. It incorporates methods used in a previous SPN structure-learning algorithm, but, unlike the previous algorithm, is not limited to learning tree-structured SPNs. Several proven ideas from circuit complexity theory along with our experimental results provide evidence for the advantages of SPNs with less-restrictive, non-tree structures.

In recommendation systems, probabilistic matrix factorization (PMF) is a state-of-the-art collaborative filtering method by determining the latent features to represent users and items. However, two major issues limiting the usefulness of PMF are the sparsity problem and long-tail distribution. Sparsity refers to the situation that the observed rating data are sparse, which results in that only part of latent features are informative for describing each item/user. Long tail distribution implies that a large fraction of items have few ratings. In this work, we propose a sparse probabilistic matrix factorization method (SPMF) by utilizing a Laplacian distribution to model the item/user factor vector. Laplacian distribution has ability to generate sparse coding, which is beneficial for SPMF to distinguish the relevant and irrelevant latent features with respect to each item/user. Meanwhile, the tails in Laplacian distribution are comparatively heavy, which is rewarding for SPMF to recommend the tail items. Furthermore, a distributed Gibbs sampling algorithm is developed to efficiently train the proposed sparse probabilistic model. A series of experiments on Netfilix and Movielens datasets have been conducted to demonstrate that SPMF outperforms the existing PMF and its extended version Bayesian PMF (BPMF), especially for the recommendation of tail items.

What does it mean for an algorithm to be biased? In U.S. law, unintentional bias is encoded via disparate impact, which occurs when a selection process has widely different outcomes for different groups, even as it appears to be neutral. This legal determination hinges on a definition of a protected class (ethnicity, gender, religious practice) and an explicit description of the process. When the process is implemented using computers, determining disparate impact (and hence bias) is harder. It might not be possible to disclose the process. In addition, even if the process is open, it might be hard to elucidate in a legal setting how the algorithm makes its decisions. Instead of requiring access to the algorithm, we propose making inferences based on the data the algorithm uses. We make four contributions to this problem. First, we link the legal notion of disparate impact to a measure of classification accuracy that while known, has received relatively little attention. Second, we propose a test for disparate impact based on analyzing the information leakage of the protected class from the other data attributes. Third, we describe methods by which data might be made unbiased. Finally, we present empirical evidence supporting the effectiveness of our test for disparate impact and our approach for both masking bias and preserving relevant information in the data. Interestingly, our approach resembles some actual selection practices that have recently received legal scrutiny.

Large multilayer neural networks trained with backpropagation have recently achieved state-of-the-art results in a wide range of problems. However, using backprop for neural net learning still has some disadvantages, e.g., having to tune a large number of hyperparameters to the data, lack of calibrated probabilistic predictions, and a tendency to overfit the training data. In principle, the Bayesian approach to learning neural networks does not have these problems. However, existing Bayesian techniques lack scalability to large dataset and network sizes. In this work we present a novel scalable method for learning Bayesian neural networks, called probabilistic backpropagation (PBP). Similar to classical backpropagation, PBP works by computing a forward propagation of probabilities through the network and then doing a backward computation of gradients. A series of experiments on ten real-world datasets show that PBP is significantly faster than other techniques, while offering competitive predictive abilities. Our experiments also show that PBP provides accurate estimates of the posterior variance on the network weights.

Gaussian processes (GP) are powerful tools for probabilistic modeling purposes. They can be used to define prior distributions over latent functions in hierarchical Bayesian models. The prior over functions is defined implicitly by the mean and covariance function, which determine the smoothness and variability of the function. The inference can then be conducted directly in the function space by evaluating or approximating the posterior process. Despite their attractive theoretical properties GPs provide practical challenges in their implementation. GPstuff is a versatile collection of computational tools for GP models compatible with Linux and Windows MATLAB and Octave. It includes, among others, various inference methods, sparse approximations and tools for model assessment. In this work, we review these tools and demonstrate the use of GPstuff in several models.

Bayesian optimization is an effective methodology for the global optimization of functions with expensive evaluations. It relies on querying a distribution over functions defined by a relatively cheap surrogate model. An accurate model for this distribution over functions is critical to the effectiveness of the approach, and is typically fit using Gaussian processes (GPs). However, since GPs scale cubically with the number of observations, it has been challenging to handle objectives whose optimization requires many evaluations, and as such, massively parallelizing the optimization. In this work, we explore the use of neural networks as an alternative to GPs to model distributions over functions. We show that performing adaptive basis function regression with a neural network as the parametric form performs competitively with state-of-the-art GP-based approaches, but scales linearly with the number of data rather than cubically. This allows us to achieve a previously intractable degree of parallelism, which we apply to large scale hyperparameter optimization, rapidly finding competitive models on benchmark object recognition tasks using convolutional networks, and image caption generation using neural language models.