The cluster assumption is exploited by most semi-supervised learning (SSL) methods. However, if the unlabeled data is merely weakly related to the target classes, it becomes questionable whether driving the decision boundary to the low density regions of the unlabeled data will help the classification. In such case, the cluster assumption may not be valid; and consequently how to leverage this type of unlabeled data to enhance the classification accuracy becomes a challenge. We introduce Semi-supervised Learning with Weakly-Related Unlabeled Data" (SSLW), an inductive method that builds upon the maximum-margin approach, towards a better usage of weakly-related unlabeled information. Although the SSLW could improve a wide range of classification tasks, in this paper, we focus on text categorization with a small training pool. The key assumption behind this work is that, even with different topics, the word usage patterns across different corpora tends to be consistent. To this end, SSLW estimates the optimal word-correlation matrix that is consistent with both the co-occurrence information derived from the weakly-related unlabeled documents and the labeled documents. For empirical evaluation, we present a direct comparison with a number of state-of-the-art methods for inductive semi-supervised learning and text categorization; and we show that SSLW results in a significant improvement in categorization accuracy, equipped with a small training set and an unlabeled resource that is weakly related to the test beds."

Much recent work in bioinformatics has focused on the inference of various types of biological networks, representing gene regulation, metabolic processes, protein-protein interactions, etc. A common setting involves inferring network edges in a supervised fashion from a set of high-confidence edges, possibly characterized by multiple, heterogeneous data sets (protein sequence, gene expression, etc.). Here, we distinguish between two modes of inference in this setting: direct inference based upon similarities between nodes joined by an edge, and indirect inference based upon similarities between one pair of nodes and another pair of nodes. We propose a supervised approach for the direct case by translating it into a distance metric learning problem. A relaxation of the resulting convex optimization problem leads to the support vector machine (SVM) algorithm with a particular kernel for pairs, which we call the metric learning pairwise kernel (MLPK). We demonstrate, using several real biological networks, that this direct approach often improves upon the state-of-the-art SVM for indirect inference with the tensor product pairwise kernel.

Predictive learning has traditionally been a standard indu ctive learning, where different sub-problem formulations have been identified. One of the most re presentative is classification, consisting on the estimation of a mapping from the feature sp ace into a finite class space. Depending on the cardinality of the finite class space we are l eft with binary or multiclass classification problems. Finally, the presence or absence o r a "natural" order among classes will separate nominal from ordinal problems. Although two-class and nominal classification problems hav e been dissected in the literature, the ordinal sibling has not yet received a lot of attention, e ven with many learning problems involving classifying examples into classes which have a na tural order. Scenarios in which it is natural to rank instances occur in many fields, such as info rmation retrieval, collaborative filtering, econometric modeling and natural sciences. Conventional methods for nominal classes or for regression problems could be employed to solve ordinal data problems; however, the use of techniques designed specifically for ordered classes yields simpler classifiers, making it easier to inte rpret the factors that are being used to discriminate among classes, and generalises better. Alt hough the ordinal formulation seems conceptually simpler than nominal, some technical di fficulties to incorporate in the algorithms this piece of additional information - the order - may explain the widespread use of conventional methods to tackle the ordinal data problem. This dissertation addresses this void by proposing a nonpar ametric procedure for the classification of ordinal data based on the extension of the original dataset with additional variables, reducing the classification task to the well-known two-clas s problem.

This paper presents a novel approach for using clickthrough data to learn ranked retrieval functions for web search results. We observe that users searching the web often perform a sequence, or chain, of queries with a similar information need. Using query chains, we generate new types of preference judgments from search engine logs, thus taking advantage of user intelligence in reformulating queries. To validate our method we perform a controlled user study comparing generated preference judgments to explicit relevance judgments. We also implemented a real-world search engine to test our approach, using a modified ranking SVM to learn an improved ranking function from preference data. Our results demonstrate significant improvements in the ranking given by the search engine. The learned rankings outperform both a static ranking function, as well as one trained without considering query chains.

Experimental physics, along with many other fields in applied and basic research, uses experiments, physical tests, and observations to gain insight into various phenomena and to validate hypotheses and models. Shock p hysics is a field that explores the response of materials to the extremes of p ressure, deformation, and temperature which are present when shock waves interact with those materials [17]. High explosive (HE) or propellant guns are often used to generate these strong shock waves. Many different diagnostic ap proaches have been used to probe these phenomena [8]. Because of the energetic nature of the shock wave drive, often a large amount of experimental equipment is destroyed during the test.

This paper deals with the problem of classifying signals. The new method for building so called local classifiers and local features is presented. The method is a combination of the lifting scheme and the support vector machines. Its main aim is to produce effective and yet comprehensible classifiers that would help in understanding processes hidden behind classified signals. To illustrate the method we present the results obtained on an artificial and a real dataset.

From a geometric perspective most nonlinear binary classification algorithms, including state of the art versions of Support Vector Machine (SVM) and Radial Basis Function Network (RBFN) classifiers, and are based on the idea of reconstructing indicator functions. We propose instead to use reconstruction of the signed distance function (SDF) as a basis for binary classification. We discuss properties of the signed distance function that can be exploited in classification algorithms. We develop simple versions of such classifiers and test them on several linear and nonlinear problems. On linear tests accuracy of the new algorithm exceeds that of standard SVM methods, with an average of 50% fewer misclassifications. Performance of the new methods also matches or exceeds that of standard methods on several nonlinear problems including classification of benchmark diagnostic micro-array data sets.

Canonical correlation analysis is a technique to extract common features from a pair of multivariate data. In complex situations, however, it does not extract useful features because of its linearity. On the other hand, kernel method used in support vector machine is an efficient approach to improve such a linear method. In this paper, we investigate the effectiveness of applying kernel method to canonical correlation analysis.

Medical Informatics and the application of modern signal processing in the assistance of the diagnostic process in medical imaging is one of the more recent and active research areas today. This thesis addresses a variety of issues related to the general problem of medical image analysis, specifically in mammography, and presents a series of algorithms and design approaches for all the intermediate levels of a modern system for computer-aided diagnosis (CAD). The diagnostic problem is analyzed with a systematic approach, first defining the imaging characteristics and features that are relevant to probable pathology in mammo-grams. Next, these features are quantified and fused into new, integrated radio-logical systems that exhibit embedded digital signal processing, in order to improve the final result and minimize the radiological dose for the patient. In a higher level, special algorithms are designed for detecting and encoding these clinically interest-ing imaging features, in order to be used as input to advanced pattern classifiers and machine learning models. Finally, these approaches are extended in multi-classifier models under the scope of Game Theory and optimum collective deci-sion, in order to produce efficient solutions for combining classifiers with minimum computational costs for advanced diagnostic systems. The material covered in this thesis is related to a total of 18 published papers, 6 in scientific journals and 12 in international conferences.

In the last few years, due to the growing ubiquity of unlabeled data, much effort has been spent by the machine learning community to develop better understanding and improve the quality of classifiers exploiting unlabeled data. Following the manifold regularization approach, Laplacian Support Vector Machines (LapSVMs) have shown the state of the art performance in semi--supervised classification. In this paper we present two strategies to solve the primal LapSVM problem, in order to overcome some issues of the original dual formulation. Whereas training a LapSVM in the dual requires two steps, using the primal form allows us to collapse training to a single step. Moreover, the computational complexity of the training algorithm is reduced from O(n^3) to O(n^2) using preconditioned conjugate gradient, where n is the combined number of labeled and unlabeled examples. We speed up training by using an early stopping strategy based on the prediction on unlabeled data or, if available, on labeled validation examples. This allows the algorithm to quickly compute approximate solutions with roughly the same classification accuracy as the optimal ones, considerably reducing the training time. Due to its simplicity, training LapSVM in the primal can be the starting point for additional enhancements of the original LapSVM formulation, such as those for dealing with large datasets. We present an extensive experimental evaluation on real world data showing the benefits of the proposed approach.