We study the regret of optimal strategies for online convex optimization games. Using von Neumann's minimax theorem, we show that the optimal regret in this adversarial setting is closely related to the behavior of the empirical minimization algorithm in a stochastic process setting: it is equal to the maximum, over joint distributions of the adversary's action sequence, of the difference between a sum of minimal expected losses and the minimal empirical loss. We show that the optimal regret has a natural geometric interpretation, since it can be viewed as the gap in Jensen's inequality for a concave functional--the minimizer over the player's actions of expected loss--defined on a set of probability distributions. We use this expression to obtain upper and lower bounds on the regret of an optimal strategy for a variety of online learning problems. Our method provides upper bounds without the need to construct a learning algorithm; the lower bounds provide explicit optimal strategies for the adversary.

For a wide variety of regularization methods, algorithms computing the entire solution path have been developed recently. Solution path algorithms do not only compute the solution for one particular value of the regularization parameter but the entire path of solutions, making the selection of an optimal parameter much easier. Most of the currently used algorithms are not robust in the sense that they cannot deal with general or degenerate input. Here we present a new robust, generic method for parametric quadratic programming. Our algorithm directly applies to nearly all machine learning applications, where so far every application required its own different algorithm. We illustrate the usefulness of our method by applying it to a very low rank problem which could not be solved by existing path tracking methods, namely to compute part-worth values in choice based conjoint analysis, a popular technique from market research to estimate consumers preferences on a class of parameterized options.

The task of identifying synonymous relations and objects, or synonym resolution, is critical for high-quality information extraction. This paper investigates synonym resolution in the context of unsupervised information extraction, where neither hand-tagged training examples nor domain knowledge is available. The paper presents a scalable, fully-implemented system that runs in O(KN log N) time in the number of extractions, N, and the maximum number of synonyms per word, K. The system, called Resolver , introduces a probabilistic relational model for predicting whether two strings are co-referential based on the similarity of the assertions containing them. On a set of two million assertions extracted from the Web, Resolver resolves objects with 78% precision and 68% recall, and resolves relations with 90% precision and 35% recall. Several variations of resolver's probabilistic model are explored, and experiments demonstrate that under appropriate conditions these variations can improve F1 by 5%. An extension to the basic Resolver system allows it to handle polysemous names with 97% precision and 95% recall on a data set from the TREC corpus.

In this work, we first show that feature selection methods other than boosting can also be used for training an efficient object detector. In particular, we introduce Greedy Sparse Linear Discriminant Analysis (GSLDA) \cite{Moghaddam2007Fast} for its conceptual simplicity and computational efficiency; and slightly better detection performance is achieved compared with \cite{Viola2004Robust}. Moreover, we propose a new technique, termed Boosted Greedy Sparse Linear Discriminant Analysis (BGSLDA), to efficiently train a detection cascade. BGSLDA exploits the sample re-weighting property of boosting and the class-separability criterion of GSLDA.

We show that the two-stage adaptive Lasso procedure (Zou, 2006) is consistent for high-dimensional model selection in linear and Gaussian graphical models. Our conditions for consistency cover more general situations than those accomplished in previous work: we prove that restricted eigenvalue conditions (Bickel et al., 2008) are also sufficient for sparse structure estimation.

We study the problem of estimating multiple linear regression equations for the purpose of both prediction and variable selection. Following recent work on multi-task learning Argyriou et al. [2008], we assume that the regression vectors share the same sparsity pattern. This means that the set of relevant predictor variables is the same across the different equations. This assumption leads us to consider the Group Lasso as a candidate estimation method. We show that this estimator enjoys nice sparsity oracle inequalities and variable selection properties. The results hold under a certain restricted eigenvalue condition and a coherence condition on the design matrix, which naturally extend recent work in Bickel et al. [2007], Lounici [2008]. In particular, in the multi-task learning scenario, in which the number of tasks can grow, we are able to remove completely the effect of the number of predictor variables in the bounds. Finally, we show how our results can be extended to more general noise distributions, of which we only require the variance to be finite.

1-Nearest Neighbor with the Dynamic Time Warping (DTW) distance is one of the most effective classifiers on time series domain. Since the global constraint has been introduced in speech community, many global constraint models have been proposed including Sakoe-Chiba (S-C) band, Itakura Parallelogram, and Ratanamahatana-Keogh (R-K) band. The R-K band is a general global constraint model that can represent any global constraints with arbitrary shape and size effectively. However, we need a good learning algorithm to discover the most suitable set of R-K bands, and the current R-K band learning algorithm still suffers from an 'overfitting' phenomenon. In this paper, we propose two new learning algorithms, i.e., band boundary extraction algorithm and iterative learning algorithm. The band boundary extraction is calculated from the bound of all possible warping paths in each class, and the iterative learning is adjusted from the original R-K band learning. We also use a Silhouette index, a well-known clustering validation technique, as a heuristic function, and the lower bound function, LB_Keogh, to enhance the prediction speed. Twenty datasets, from the Workshop and Challenge on Time Series Classification, held in conjunction of the SIGKDD 2007, are used to evaluate our approach.

For tensor decompositions such as HOSVD and ParaFac, the objective functions are nonconvex. This implies, theoretically, there exists a large number of local optimas: starting from different starting point, the iteratively improved solution will converge to different local solutions. This non-uniqueness present a stability and reliability problem for image compression and retrieval. In this paper, we present the results of a comprehensive investigation of this problem. We found that although all tensor decomposition algorithms fail to reach a unique global solution on random data and severely scrambled data; surprisingly however, on all real life several data sets (even with substantial scramble and occlusions), HOSVD always produce the unique global solution in the parameter region suitable to practical applications, while ParaFac produce non-unique solutions. We provide an eigenvalue based rule for the assessing the solution uniqueness.

This paper addresses the general problem of domain adaptation which arises in a variety of applications where the distribution of the labeled sample available somewhat differs from that of the test data. Building on previous work by Ben-David et al. (2007), we introduce a novel distance between distributions, discrepancy distance, that is tailored to adaptation problems with arbitrary loss functions. We give Rademacher complexity bounds for estimating the discrepancy distance from finite samples for different loss functions. Using this distance, we derive novel generalization bounds for domain adaptation for a wide family of loss functions. We also present a series of novel adaptation bounds for large classes of regularization-based algorithms, including support vector machines and kernel ridge regression based on the empirical discrepancy. This motivates our analysis of the problem of minimizing the empirical discrepancy for various loss functions for which we also give novel algorithms. We report the results of preliminary experiments that demonstrate the benefits of our discrepancy minimization algorithms for domain adaptation.

The runtime for Kernel Partial Least Squares (KPLS) to compute the fit is quadratic in the number of examples. However, the necessity of obtaining sensitivity measures as degrees of freedom for model selection or confidence intervals for more detailed analysis requires cubic runtime, and thus constitutes a computational bottleneck in real-world data analysis. We propose a novel algorithm for KPLS which not only computes (a) the fit, but also (b) its approximate degrees of freedom and (c) error bars in quadratic runtime. The algorithm exploits a close connection between Kernel PLS and the Lanczos algorithm for approximating the eigenvalues of symmetric matrices, and uses this approximation to compute the trace of powers of the kernel matrix in quadratic runtime.