This paper considers the multi-task learning problem and in the setting where some relevant features could be shared across few related tasks. Most of the existing methods assume the extent to which the given tasks are related or share a common feature space to be known apriori. In real-world applications however, it is desirable to automatically discover the groups of related tasks that share a feature space. In this paper we aim at searching the exponentially large space of all possible groups of tasks that may share a feature space. The main contribution is a convex formulation that employs a graph-based regularizer and simultaneously discovers few groups of related tasks, having close-by task parameters, as well as the feature space shared within each group. The regularizer encodes an important structure among the groups of tasks leading to an efficient algorithm for solving it: if there is no feature space under which a group of tasks has close-by task parameters, then there does not exist such a feature space for any of its supersets. An efficient active set algorithm that exploits this simplification and performs a clever search in the exponentially large space is presented. The algorithm is guaranteed to solve the proposed formulation (within some precision) in a time polynomial in the number of groups of related tasks discovered. Empirical results on benchmark datasets show that the proposed formulation achieves good generalization and outperforms state-of-the-art multi-task learning algorithms in some cases.

In this paper we present a fully Bayesian latent variable model which exploits conditional nonlinear(in)-dependence structures to learn an efficient latent representation. The latent space is factorized to represent shared and private information from multiple views of the data. In contrast to previous approaches, we introduce a relaxation to the discrete segmentation and allow for a "softly" shared latent space. Further, Bayesian techniques allow us to automatically estimate the dimensionality of the latent spaces. The model is capable of capturing structure underlying extremely high dimensional spaces. This is illustrated by modelling unprocessed images with tenths of thousands of pixels. This also allows us to directly generate novel images from the trained model by sampling from the discovered latent spaces. We also demonstrate the model by prediction of human pose in an ambiguous setting. Our Bayesian framework allows us to perform disambiguation in a principled manner by including latent space priors which incorporate the dynamic nature of the data.

We consider the most common variants of linear regression, including Ridge, Lasso and Support-vector regression, in a setting where the learner is allowed to observe only a fixed number of attributes of each example at training time. We present simple and efficient algorithms for these problems: for Lasso and Ridge regression they need the same total number of attributes (up to constants) as do full-information algorithms, for reaching a certain accuracy. For Support-vector regression, we require exponentially less attributes compared to the state of the art. By that, we resolve an open problem recently posed by Cesa-Bianchi et al. (2010). Experiments show the theoretical bounds to be justified by superior performance compared to the state of the art.

Algorithms for learning distributions over weight-vectors, such as AROW were recently shown empirically to achieve state-of-the-art performance at various problems, with strong theoretical guaranties. Extending these algorithms to matrix models pose challenges since the number of free parameters in the covariance of the distribution scales as $n^4$ with the dimension $n$ of the matrix, and $n$ tends to be large in real applications. We describe, analyze and experiment with two new algorithms for learning distribution of matrix models. Our first algorithm maintains a diagonal covariance over the parameters and can handle large covariance matrices. The second algorithm factors the covariance to capture inter-features correlation while keeping the number of parameters linear in the size of the original matrix. We analyze both algorithms in the mistake bound model and show a superior precision performance of our approach over other algorithms in two tasks: retrieving similar images, and ranking similar documents. The factored algorithm is shown to attain faster convergence rate.

Retrieval tasks typically require a ranking of items given a query. Collaborative filtering tasks, on the other hand, learn to model user's preferences over items. In this paper we study the joint problem of recommending items to a user with respect to a given query, which is a surprisingly common task. This setup differs from the standard collaborative filtering one in that we are given a query x user x item tensor for training instead of the more traditional user x item matrix. Compared to document retrieval we do have a query, but we may or may not have content features (we will consider both cases) and we can also take account of the user's profile. We introduce a factorized model for this new task that optimizes the top-ranked items returned for the given query and user. We report empirical results where it outperforms several baselines.

We consider the problem of parameter estimation using weakly supervised datasets, where a training sample consists of the input and a partially specified annotation, which we refer to as the output. The missing information in the annotation is modeled using latent variables. Previous methods overburden a single distribution with two separate tasks: (i) modeling the uncertainty in the latent variables during training; and (ii) making accurate predictions for the output and the latent variables during testing. We propose a novel framework that separates the demands of the two tasks using two distributions: (i) a conditional distribution to model the uncertainty of the latent variables for a given input-output pair; and (ii) a delta distribution to predict the output and the latent variables for a given input. During learning, we encourage agreement between the two distributions by minimizing a loss-based dissimilarity coefficient. Our approach generalizes latent SVM in two important ways: (i) it models the uncertainty over latent variables instead of relying on a pointwise estimate; and (ii) it allows the use of loss functions that depend on latent variables, which greatly increases its applicability. We demonstrate the efficacy of our approach on two challenging problems---object detection and action detection---using publicly available datasets.

Positive definite operator-valued kernels generalize the well-known notion of reproducing kernels, and are naturally adapted to multi-output learning situations. This paper addresses the problem of learning a finite linear combination of infinite-dimensional operator-valued kernels which are suitable for extending functional data analysis methods to nonlinear contexts. We study this problem in the case of kernel ridge regression for functional responses with an lr-norm constraint on the combination coefficients. The resulting optimization problem is more involved than those of multiple scalar-valued kernel learning since operator-valued kernels pose more technical and theoretical issues. We propose a multiple operator-valued kernel learning algorithm based on solving a system of linear operator equations by using a block coordinatedescent procedure. We experimentally validate our approach on a functional regression task in the context of finger movement prediction in brain-computer interfaces.

This manuscript studies statistical properties of linear classifiers obtained through minimization of an unregularized convex risk over a finite sample. Although the results are explicitly finite-dimensional, inputs may be passed through feature maps; in this way, in addition to treating the consistency of logistic regression, this analysis also handles boosting over a finite weak learning class with, for instance, the exponential, logistic, and hinge losses. In this finite-dimensional setting, it is still possible to fit arbitrary decision boundaries: scaling the complexity of the weak learning class with the sample size leads to the optimal classification risk almost surely.

We propose a variable decomposition algorithm -greedy block coordinate descent (GBCD)- in order to make dense Gaussian process regression practical for large scale problems. GBCD breaks a large scale optimization into a series of small sub-problems. The challenge in variable decomposition algorithms is the identification of a subproblem (the active set of variables) that yields the largest improvement. We analyze the limitations of existing methods and cast the active set selection into a zero-norm constrained optimization problem that we solve using greedy methods. By directly estimating the decrease in the objective function, we obtain not only efficient approximate solutions for GBCD, but we are also able to demonstrate that the method is globally convergent. Empirical comparisons against competing dense methods like Conjugate Gradient or SMO show that GBCD is an order of magnitude faster. Comparisons against sparse GP methods show that GBCD is both accurate and capable of handling datasets of 100,000 samples or more.

Exemplar-based clustering methods have been shown to produce state-of-the-art results on a number of synthetic and real-world clustering problems. They are appealing because they offer computational benefits over latent-mean models and can handle arbitrary pairwise similarity measures between data points. However, when trying to recover underlying structure in clustering problems, tailored similarity measures are often not enough; we also desire control over the distribution of cluster sizes. Priors such as Dirichlet process priors allow the number of clusters to be unspecified while expressing priors over data partitions. To our knowledge, they have not been applied to exemplar-based models. We show how to incorporate priors, including Dirichlet process priors, into the recently introduced affinity propagation algorithm. We develop an efficient maxproduct belief propagation algorithm for our new model and demonstrate experimentally how the expanded range of clustering priors allows us to better recover true clusterings in situations where we have some information about the generating process.