Black-box alpha (BB-$\alpha$) is a new approximate inference method based on the minimization of $\alpha$-divergences. BB-$\alpha$ scales to large datasets because it can be implemented using stochastic gradient descent. BB-$\alpha$ can be applied to complex probabilistic models with little effort since it only requires as input the likelihood function and its gradients. These gradients can be easily obtained using automatic differentiation. By changing the divergence parameter $\alpha$, the method is able to interpolate between variational Bayes (VB) ($\alpha \rightarrow 0$) and an algorithm similar to expectation propagation (EP) ($\alpha = 1$). Experiments on probit regression and neural network regression and classification problems show that BB-$\alpha$ with non-standard settings of $\alpha$, such as $\alpha = 0.5$, usually produces better predictions than with $\alpha \rightarrow 0$ (VB) or $\alpha = 1$ (EP).

This paper considers the recovery of a rank $r$ positive semidefinite matrix $X X^T\in\mathbb{R}^{n\times n}$ from $m$ scalar measurements of the form $y_i := a_i^T X X^T a_i$ (i.e., quadratic measurements of $X$). Such problems arise in a variety of applications, including covariance sketching of high-dimensional data streams, quadratic regression, quantum state tomography, among others. A natural approach to this problem is to minimize the loss function $f(U) = \sum_i (y_i - a_i^TUU^Ta_i)^2$ which has an entire manifold of solutions given by $\{XO\}_{O\in\mathcal{O}_r}$ where $\mathcal{O}_r$ is the orthogonal group of $r\times r$ orthogonal matrices; this is {\it non-convex} in the $n\times r$ matrix $U$, but methods like gradient descent are simple and easy to implement (as compared to semidefinite relaxation approaches). In this paper we show that once we have $m \geq C nr \log^2(n)$ samples from isotropic gaussian $a_i$, with high probability {\em (a)} this function admits a dimension-independent region of {\em local strong convexity} on lines perpendicular to the solution manifold, and {\em (b)} with an additional polynomial factor of $r$ samples, a simple spectral initialization will land within the region of convexity with high probability. Together, this implies that gradient descent with initialization (but no re-sampling) will converge linearly to the correct $X$, up to an orthogonal transformation. We believe that this general technique (local convexity reachable by spectral initialization) should prove applicable to a broader class of nonconvex optimization problems.

Last week ago I posted an article called, What Do Predictive Analytics Consultants Do? Part 1, describing the general types of activities that we engage in. In the present article, I want to talk about the skills and tools that one should have to perform Predictive Analytics. Although this is not strictly a "What we do" article, knowing the skills we possess and the tools we use will provide some insight into what we do, without talking about some algorithm that you may have never heard of. I am always at a loss in describing the skills of analytics, for there are many. I just completed a new book about analytics (available for FREE--see notes) that has a different approach than Predictive Analytics using R (also available for FREE), though I am using material from three chapters.

In the latest post of my own blog, I argued about how to learn how each machine learning classifier works visually. My idea is that first I prepare samples for training and then I show its assumed true boundary, and finally decision boundary estimated by the classifier with a dense grid covering over the space as test dataset and the assumed boundary are compared. In the case below, the assumed true boundary of the space is a set of 3 parallel lines; I think everybody will guess so intuitively, but the most important point here is whether any machine learning classifier works so. For example, when multinomial logit - one of linear classifiers - is trained by samples below, it gives decision boundary for a grid dataset covering the whole space. It looks almost the same as the assumed boundary.

Classical collaborative filtering, and content-based filtering methods try to learn a static recommendation model given training data. These approaches are far from ideal in highly dynamic recommendation domains such as news recommendation and computational advertisement, where the set of items and users is very fluid. In this work, we investigate an adaptive clustering technique for content recommendation based on exploration-exploitation strategies in contextual multi-armed bandit settings. Our algorithm takes into account the collaborative effects that arise due to the interaction of the users with the items, by dynamically grouping users based on the items under consideration and, at the same time, grouping items based on the similarity of the clusterings induced over the users. The resulting algorithm thus takes advantage of preference patterns in the data in a way akin to collaborative filtering methods. We provide an empirical analysis on medium-size real-world datasets, showing scalability and increased prediction performance (as measured by click-through rate) over state-of-the-art methods for clustering bandits. We also provide a regret analysis within a standard linear stochastic noise setting.

Mixture proportion estimation (MPE) is the problem of estimating the weight of a component distribution in a mixture, given samples from the mixture and component. This problem constitutes a key part in many "weakly supervised learning" problems like learning with positive and unlabelled samples, learning with label noise, anomaly detection and crowdsourcing. While there have been several methods proposed to solve this problem, to the best of our knowledge no efficient algorithm with a proven convergence rate towards the true proportion exists for this problem. We fill this gap by constructing a provably correct algorithm for MPE, and derive convergence rates under certain assumptions on the distribution. Our method is based on embedding distributions onto an RKHS, and implementing it only requires solving a simple convex quadratic programming problem a few times. We run our algorithm on several standard classification datasets, and demonstrate that it performs comparably to or better than other algorithms on most datasets.

We present CYCLADES, a general framework for parallelizing stochastic optimization algorithms in a shared memory setting. CYCLADES is asynchronous during shared model updates, and requires no memory locking mechanisms, similar to HOGWILD!-type algorithms. Unlike HOGWILD!, CYCLADES introduces no conflicts during the parallel execution, and offers a black-box analysis for provable speedups across a large family of algorithms. Due to its inherent conflict-free nature and cache locality, our multi-core implementation of CYCLADES consistently outperforms HOGWILD!-type algorithms on sufficiently sparse datasets, leading to up to 40% speedup gains compared to the HOGWILD! implementation of SGD, and up to 5x gains over asynchronous implementations of variance reduction algorithms.

The restricted isometry property (RIP) for design matrices gives guarantees for optimal recovery in sparse linear models. It is of high interest in compressed sensing and statistical learning. This property is particularly important for computationally efficient recovery methods. As a consequence, even though it is in general NP-hard to check that RIP holds, there have been substantial efforts to find tractable proxies for it. These would allow the construction of RIP matrices and the polynomial-time verification of RIP given an arbitrary matrix. We consider the framework of average-case certifiers, that never wrongly declare that a matrix is RIP, while being often correct for random instances. While there are such functions which are tractable in a suboptimal parameter regime, we show that this is a computationally hard task in any better regime. Our results are based on a new, weaker assumption on the problem of detecting dense subgraphs.

Given a symmetric nonnegative matrix $A$, symmetric nonnegative matrix factorization (symNMF) is the problem of finding a nonnegative matrix $H$, usually with much fewer columns than $A$, such that $A \approx HH^T$. SymNMF can be used for data analysis and in particular for various clustering tasks. In this paper, we propose simple and very efficient coordinate descent schemes to solve this problem, and that can handle large and sparse input matrices. The effectiveness of our methods is illustrated on synthetic and real-world data sets, and we show that they perform favorably compared to recent state-of-the-art methods.

Machine learning automatically recognizes complex, previously unknown, novel, and useful patterns and information in all types of data. Data driven algorithms are the wave of the future and their results improve as the amount of data increases. Machine learning algorithms are used in search engines, image analysis, multimedia database retrieval, bioinformatics, industrial automation, speech recognition, and many other fields. This survey course covers the concepts and principles of a large variety of data mining methods, equips you with a working knowledge of these techniques and prepares you to apply them to real problems. The statistical programming language R is used to implement machine learning algorithms.