There has recently been considerable interest in completing a low-rank matrix or tensor given only a small fraction (or few linear combinations) of its entries. Related approaches have found considerable success in the area of recommender systems, under machine learning. From a statistical estimation point of view, the gold standard is to have access to the joint probability distribution of all pertinent random variables, from which any desired optimal estimator can be readily derived. In practice high-dimensional joint distributions are very hard to estimate, and only estimates of low-dimensional projections may be available. We show that it is possible to identify higher-order joint PMFs from lower-order marginalized PMFs using coupled low-rank tensor factorization. Our approach features guaranteed identifiability when the full joint PMF is of low-enough rank, and effective approximation otherwise. We provide an algorithmic approach to compute the sought factors, and illustrate the merits of our approach using rating prediction as an example.

Adaptive sampling theory has shown that, with proper assumptions on the signal class, algorithms exist to reconstruct a signal in $\mathbb{R}^{d}$ with an optimal number of samples. We generalize this problem to the case of spatial signals, where the sampling cost is a function of both the number of samples taken and the distance traveled during estimation. This is motivated by our work studying regions of low oxygen concentration in the Great Lakes. We show that for one-dimensional threshold classifiers, a tradeoff between the number of samples taken and distance traveled can be achieved using a generalization of binary search, which we refer to as quantile search. We characterize both the estimation error after a fixed number of samples and the distance traveled in the noiseless case, as well as the estimation error in the case of noisy measurements. We illustrate our results in both simulations and experiments and show that our method outperforms existing algorithms in the majority of practical scenarios.

In the last few months, I have had several people contact me about their enthusiasm for venturing into the world of data science and using Machine Learning (ML) techniques to probe statistical regularities and build impeccable data-driven products. However, I've observed that some actually lack the necessary mathematical intuition and framework to get useful results. This is the main reason I decided to write this blog post. Recently, there has been an upsurge in the availability of many easy-to-use machine and deep learning packages such as scikit-learn, Weka, Tensorflow etc. Machine Learning theory is a field that intersects statistical, probabilistic, computer science and algorithmic aspects arising from learning iteratively from data and finding hidden insights which can be used to build intelligent applications. Despite the immense possibilities of Machine and Deep Learning, a thorough mathematical understanding of many of these techniques is necessary for a good grasp of the inner workings of the algorithms and getting good results.

Multi-label classification is a type of supervised learning where an instance may belong to multiple labels simultaneously. Predicting each label independently has been criticized for not exploiting any correlation between labels. In this paper we propose a novel approach, Nearest Labelset using Double Distances (NLDD), that predicts the labelset observed in the training data that minimizes a weighted sum of the distances in both the feature space and the label space to the new instance. The weights specify the relative tradeoff between the two distances. The weights are estimated from a binomial regression of the number of misclassified labels as a function of the two distances. Model parameters are estimated by maximum likelihood. NLDD only considers labelsets observed in the training data, thus implicitly taking into account label dependencies. Experiments on benchmark multi-label data sets show that the proposed method on average outperforms other well-known approaches in terms of Hamming loss, 0/1 loss, and multi-label accuracy and ranks second after ECC on the F-measure.

The reconstruction of an object's shape or surface from a set of 3D points plays an important role in medical image analysis, e.g. in anatomy reconstruction from tomographic measurements or in the process of aligning intra-operative navigation and preoperative planning data. In such scenarios, one usually has to deal with sparse data, which significantly aggravates the problem of reconstruction. However, medical applications often provide contextual information about the 3D point data that allow to incorporate prior knowledge about the shape that is to be reconstructed. To this end, we propose the use of a statistical shape model (SSM) as a prior for surface reconstruction. The SSM is represented by a point distribution model (PDM), which is associated with a surface mesh. Using the shape distribution that is modelled by the PDM, we formulate the problem of surface reconstruction from a probabilistic perspective based on a Gaussian Mixture Model (GMM). In order to do so, the given points are interpreted as samples of the GMM. By using mixture components with anisotropic covariances that are "oriented" according to the surface normals at the PDM points, a surface-based fitting is accomplished. Estimating the parameters of the GMM in a maximum a posteriori manner yields the reconstruction of the surface from the given data points. We compare our method to the extensively used Iterative Closest Points method on several different anatomical datasets/SSMs (brain, femur, tibia, hip, liver) and demonstrate superior accuracy and robustness on sparse data.

Getting learners to read textbooks and use other teaching aids effectively can be tricky. Especially, when the books are just too dreary. In this post, we've compiled great e-resources for you digital natives looking to explore the exciting world of Machine Learning and Neural Networks. But before you dive into the deep end, you need to make sure you've got the fundamentals down pat. It doesn't matter what catches your fancy, machine learning, artificial intelligence, or deep learning; you need to know the basics of math and stats--linear algebra, calculus, optimization, probability--to get ahead.

Unsupervised neural networks, such as restricted Boltzmann machines (RBMs) and deep belief networks (DBNs), are powerful tools for feature selection and pattern recognition tasks. We demonstrate that overfitting occurs in such models just as in deep feedforward neural networks, and discuss possible regularization methods to reduce overfitting. We also propose a "partial" approach to improve the efficiency of Dropout/DropConnect in this scenario, and discuss the theoretical justification of these methods from model convergence and likelihood bounds. Finally, we compare the performance of these methods based on their likelihood and classification error rates for various pattern recognition data sets.

Multivariate count data are pervasive in science in the form of histograms, contingency tables and others. Previous work on modeling this type of distributions do not allow for fast and tractable inference. In this paper we present a novel Poisson graphical model, the first based on sum product networks, called PSPN, allowing for positive as well as negative dependencies. We present algorithms for learning tree PSPNs from data as well as for tractable inference via symbolic evaluation. With these, information-theoretic measures such as entropy, mutual information, and distances among count variables can be computed without resorting to approximations. Additionally, we show a connection between PSPNs and LDA, linking the structure of tree PSPNs to a hierarchy of topics. The experimental results on several synthetic and real world datasets demonstrate that PSPN often outperform state-of-the-art while remaining tractable.

Weighted model counting (WMC) has recently emerged as an effective and general approach to probabilistic inference, offering a computational framework for encoding a variety of formalisms, such as factor graphs and Bayesian networks.The advent of large-scale probabilistic knowledge bases has generated further interest in relational probabilistic representations, obtained by according weights to first-order formulas, whose semantics is given in terms of the ground theory, and solved by WMC. A fundamental limitation is that the domain of quantification, by construction and design, is assumed to be finite, which is at odds with areas such as vision and language understanding, where the existence of objects must be inferred from raw data. Dropping the finite-domain assumption has been known to improve the expressiveness of a first-order language for open-universe purposes, but these languages, so far, have eluded WMC approaches. In this paper, we revisit relational probabilistic models over an infinite domain, and establish a number of results that permit effective algorithms. We demonstrate this language on a number of examples, including a parameterized version of Pearl's Burglary-Earthquake-Alarm Bayesian network.

Several advanced applications of autonomous aerial vehicles in civilian and military contexts involve a searching agent with imperfect sensors that seeks to locate a mobile target in a given region. Effectively managing uncertainty is key to solving the related search problem, which is why all methods devised so far hinge on a probabilistic formulation of the problem and solve it through branch-and-bound algorithms, Bayesian filtering or POMDP solvers. In this paper, we consider a class of hard search tasks involving a target that exhibits an intentional evasive behaviour and moves over a large geographical area, i.e., a target that is particularly difficult to track down and uncertain to locate. We show that, even for such a complex problem, it is advantageous to compile its probabilistic structure into a deterministic model and use standard deterministic solvers to find solutions. In particular, we formulate the search problem for our uncooperative target both as a deterministic automated planning task and as a constraint programming task and show that in both cases our solution outperforms POMDPs methods.