Recommendation systems are an increasingly prominent part of the web, accounting for up to a third of all traffic on several of the world’s most popular sites. Nevertheless, little is known about how much activity such systems actually cause over and above activity that would have occurred via other means (e.g., search) if recommendations were absent. Although the ideal way to estimate the causal impact of recommendations is via randomized experiments, such experiments are costly and may inconvenience users. In this paper, therefore, we present a method for estimating causal effects from purely observational data. Specifically, we show that causal identification through an instrumental variable is possible when a product experiences an instantaneous shock in direct traffic and the products recommended next to it do not. We then apply our method to browsing logs containing anonymized activity for 2.1 million users on Amazon.com over a 9 month period and analyze over 4,000 unique products that experience such shocks. We find that although recommendation click-throughs do account for a large fraction of traffic among these products, at least 75% of this activity would likely occur in the absence of recommendations. We conclude with a discussion about the assumptions under which the method is appropriate and caveats around extrapolating results to other products, sites, or settings.

Given a network with attributed edges, how can we identify anomalous behavior? Networks with edge attributes are commonplace in the real world. For example, edges in e-commerce networks often indicate how users rated products and services in terms of number of stars, and edges in online social and phonecall networks contain temporal information about when friendships were formed and when users communicated with each other — in such cases, edge attributes capture information about how the adjacent nodes interact with other entities in the network. In this paper, we aim to utilize exactly this information to discern suspicious from typical node behavior. Our work has a number of notable contributions, including (a) formulation: while most other graph-based anomaly detection works use structural graph connectivity or node information, we focus on the new problem of leveraging edge information, (b) methodology: we introduce EdgeCentric, an intuitive and scalable compression-based approach for detecting edge-attributed graph anomalies, and (c) practicality: we show that EdgeCentric successfully spots numerous such anomalies in several large, edge-attributed real-world graphs, including the Flipkart e-commerce graph with over 3 million product reviews between 1.1 million users and 545 thousand products, where it achieved 0.87 precision over the top 100 results.

In this work we study preprocessing for tractable problems when part of the input is unknown or uncertain. This comes up naturally if, e.g., the load of some machines or the congestion of some roads is not known far enough in advance, or if we have to regularly solve a problem over instances that are largely similar, e.g., daily airport scheduling with few charter flights. Unlike robust optimization, which also studies settings like this, our goal lies not in computing solutions that are (approximately) good for every instantiation. Rather, we seek to preprocess the known parts of the input, to speed up finding an optimal solution once the missing data is known. We present efficient algorithms that given an instance with partially uncertain input generate an instance of size polynomial in the amount of uncertain data that is equivalent for every instantiation of the unknown part. Concretely, we obtain such algorithms for Minimum Spanning Tree, Minimum Weight Matroid Basis, and Maximum Cardinality Bipartite Maxing, where respectively the weight of edges, weight of elements, and the availability of vertices is unknown for part of the input. Furthermore, we show that there are tractable problems, such as Small Connected Vertex Cover, for which one cannot hope to obtain similar results.

In this paper the exact linear relation between the leading eigenvectors of the modularity matrix and the singular vectors of an uncentered data matrix is developed. Based on this analysis the concept of a modularity component is defined, and its properties are developed. It is shown that modularity component analysis can be used to cluster data similar to how traditional principal component analysis is used except that modularity component analysis does not require data centering.

Clustering algorithms start with a fixed divergence metric, which captures the possibly asymmetric distance between two samples. In a mixture model, the sample distribution plays the role of a divergence metric. It is often the case that the distributional assumption is not validated, which calls for an adaptive approach. We consider a richer model where the underlying distribution belongs to a parametrized exponential family, called Tweedie Models. We first show the connection between the Tweedie models and beta divergences, and derive the corresponding hard-assignment clustering algorithm. We exploit this connection to identify moment conditions and use Generalized Method of Moments(GMoM) to learn the data distribution. Based on this adaptive approach, we propose four new hard clustering algorithms and compare them to the classical k-means and DP-means on synthetic data as well as seven UCI datasets and one large gene expression dataset. We further compare the GMoM routine to an approximate maximum likelihood routine and validate the computational benefits of the GMoM approach.

It is well known that most of the common clustering objectives are NP-hard to optimize. In practice, however, clustering is being routinely carried out. One approach for providing theoretical understanding of this seeming discrepancy is to come up with notions of clusterability that distinguish realistically interesting input data from worst-case data sets. The hope is that there will be clustering algorithms that are provably efficient on such ‘clusterable’ instances. This paper addresses the thesis that the computational hardness of clustering tasks goes away for inputs that one really cares about. In other words, that ‘Clustering is difficult only when it does not matter’ (the \emph{CDNM thesis} for short). I wish to present a a critical bird’s eye overview of the results published on this issue so far and to call attention to the gap between available and desirable results on this issue. A longer, more detailed version of this note is available as arXiv:1507.05307. I discuss which requirements should be met in order to provide formal support to the the CDNM thesis and then examine existing results in view of these requirements and list some significant unsolved research challenges in that direction.

With the emergence of social networking services, researchers enjoy the increasing availability of large-scale heterogenous datasets capturing online user interactions and behaviors. Traditional analysis of techno-social systems data has focused mainly on describing either the dynamics of social interactions, or the attributes and behaviors of the users. However, overwhelming empirical evidence suggests that the two dimensions affect one another, and therefore they should be jointly modeled and analyzed in a multi-modal framework. The benefits of such an approach include the ability to build better predictive models, leveraging social network information as well as user behavioral signals. To this purpose, here we propose the Constrained Latent Space Model (CLSM), a generalized framework that combines Mixed Membership Stochastic Blockmodels (MMSB) and Latent Dirichlet Allocation (LDA) incorporating a constraint that forces the latent space to concurrently describe the multiple data modalities. We derive an efficient inference algorithm based on Variational Expectation Maximization that has a computational cost linear in the size of the network, thus making it feasible to analyze massive social datasets. We validate the proposed framework on two problems: prediction of social interactions from user attributes and behaviors, and behavior prediction exploiting network information. We perform experiments with a variety of multi-modal social systems, spanning location-based social networks (Gowalla), social media services (Instagram, Orkut), e-commerce and review sites (Amazon, Ciao), and finally citation networks (Cora). The results indicate significant improvement in prediction accuracy over state of the art methods, and demonstrate the flexibility of the proposed approach for addressing a variety of different learning problems commonly occurring with multi-modal social data.

The growing incidents of counterfeiting and associated economic and health consequences necessitate the development of active surveillance systems capable of producing timely and reliable information for all stake holders in the anti-counterfeiting fight. User generated content from social media platforms can provide early clues about product allergies, adverse events and product counterfeiting. This paper reports a work in progresswith contributions including: the development of a framework for gathering and analyzing the views and experiences of users of drug and cosmetic products using machine learning, text mining and sentiment analysis, the application of the proposed framework on Facebook comments and data from Twitter for brand analysis, and the description of how to develop a product safety lexicon and training data for modeling a machine learning classifier for drug and cosmetic product sentiment prediction. The initial brand and product comparison results signify the usefulness of text mining and sentiment analysis on social media data while the use of machine learning classifier for predicting the sentiment orientation provides a useful tool for users, product manufacturers, regulatory and enforcement agencies to monitor brand or product sentiment trends in order to act in the event of sudden or significant rise in negative sentiment.

This paper reviews the design of experiments for generalised linear models, including optimal design, Bayesian design and designs for models with random effects.

Non-negative matrix factorization (NMF) is a common method for generating topic models from text data. NMF is widely accepted for producing good results despite its relative simplicity of implementation and ease of computation. One challenge with applying NMF to large datasets is that intermediate matrix products often become dense, stressing the memory and compute elements of a system. In this article, we investigate a simple but powerful modification of a common NMF algorithm that enforces the generation of sparse intermediate and output matrices. This method enables the application of NMF to large datasets through improved memory and compute performance. Further, we demonstrate empirically that this method of enforcing sparsity in the NMF either preserves or improves both the accuracy of the resulting topic model and the convergence rate of the underlying algorithm.

A causal falling rule list (CFRL) is a sequence of if-then rules that specifies heterogeneous treatment effects, where (i) the order of rules determines the treatment effect subgroup a subject belongs to, and (ii) the treatment effect decreases monotonically down the list. A given CFRL parameterizes a hierarchical bayesian regression model in which the treatment effects are incorporated as parameters, and assumed constant within model-specific subgroups. We formulate the search for the CFRL best supported by the data as a Bayesian model selection problem, where we perform a search over the space of CFRL models, and approximate the evidence for a given CFRL model using standard variational techniques. We apply CFRL to a census wage dataset to identify subgroups of differing wage inequalities between men and women.

We demonstrate how to use simple dynamical processes such as random walks to study and visualize community structure in multilayer networks.

Robust Bayesian models are appealing alternatives to standard models, providing protection from data that contains outliers or other departures from the model assumptions. Historically, robust models were mostly developed on a case-by-case basis; examples include robust linear regression, robust mixture models, and bursty topic models. In this paper we develop a general approach to robust Bayesian modeling. We show how to turn an existing Bayesian model into a robust model, and then develop a generic strategy for computing with it. We use our method to study robust variants of several models, including linear regression, Poisson regression, logistic regression, and probabilistic topic models. We discuss the connections between our methods and existing approaches, especially empirical Bayes and James-Stein estimation.

Basic Linear Algebra Subprograms (BLAS) are a set of low level linear algebra kernels widely adopted by applications involved with the deep learning and scientific computing. The massive and economic computing power brought forth by the emerging GPU architectures drives interest in implementation of compute-intensive level 3 BLAS on multi-GPU systems. In this paper, we investigate existing multi-GPU level 3 BLAS and present that 1) issues, such as the improper load balancing, inefficient communication, insufficient GPU stream level concurrency and data caching, impede current implementations from fully harnessing heterogeneous computing resources; 2) and the inter-GPU Peer-to-Peer(P2P) communication remains unexplored. We then present BLASX: a highly optimized multi-GPU level-3 BLAS. We adopt the concepts of algorithms-by-tiles treating a matrix tile as the basic data unit and operations on tiles as the basic task. Tasks are guided with a dynamic asynchronous runtime, which is cache and locality aware. The communication cost under BLASX becomes trivial as it perfectly overlaps communication and computation across multiple streams during asynchronous task progression. It also takes the current tile cache scheme one step further by proposing an innovative 2-level hierarchical tile cache, taking advantage of inter-GPU P2P communication. As a result, linear speedup is observable with BLASX under multi-GPU configurations; and the extensive benchmarks demonstrate that BLASX consistently outperforms the related leading industrial and academic projects such as cuBLAS-XT, SuperMatrix, MAGMA and PaRSEC.

This is the first of a series of papers that the authors propose to write on the subject of improving the speed of response of learning systems using multiple models. During the past two decades, the second author has worked on numerous methods for improving the stability, robustness, and performance of adaptive systems using multiple models and the other authors have collaborated with him on some of them. Independently, they have also worked on several learning methods, and have considerable experience with their advantages and limitations. In particular, they are well aware that it is common knowledge that machine learning is in general very slow. Numerous attempts have been made by researchers to improve the speed of convergence of algorithms in different contexts. In view of the success of multiple model based methods in improving the speed of convergence in adaptive systems, the authors believe that the same approach will also prove fruitful in the domain of learning. In this paper, a first attempt is made to use multiple models for improving the speed of response of the simplest learning schemes that have been studied. i.e. Learning Automata.