One goal of research in artificial intelligence is to automate tasks that currently require human expertise; this automation is important because it saves time and brings problems that were previously too large to be solved into the feasible domain. Data analysis, or the ability to identify meaningful patterns and trends in large volumes of data, is an important task that falls into this category. Clustering algorithms are a particularly useful group of data analysis tools. These methods are used, for example, to analyze satellite images of the Earth to identify and categorize different land and foliage types or to analyze telescopic observations to determine what distinct types of astronomical bodies exist and to categorize each observation. However, most existing clustering methods apply general similarity techniques rather than making use of problem-specific information. This dissertation first presents a novel method for converting existing clustering algorithms into constrained clustering algorithms. The resulting methods are able to accept domain-specific information in the form of constraints on the output clusters. At the most general level, each constraint is an instance-level statement

Semi-supervised clustering uses a small amount of labeled data to aid and bias the clustering of unlabeled data. This paper explores the use of labeled data to generate initial seed clusters, as well as the use of constraints generated from labeled data to guide the clustering process. It introduces two semi-supervised variants of KMeans clustering that can be viewed as instances of the EM algorithm, where labeled data provides prior information about the conditional distributions of hidden category labels. Experimental results demonstrate the advantages of these methods over standard random seeding and COP-KMeans, a previously developed semi-supervised clustering algorithm.

We present a general framework for discriminative estimation based on the maximum entropy principle and its extensions. All calculations involve distributions over structures and/or parameters rather than specific settings and reduce to relative entropy projections. This holds even when the data is not separable within the chosen parametric class, in the context of anomaly detection rather than classification, or when the labels in the training set are uncertain or incomplete. Support vector machines are naturally subsumed under this class and we provide several extensions. We are also able to estimate exactly and efficiently discriminative distributions over tree structures of class-conditional models within this framework. Preliminary experimental results are indicative of the potential in these techniques.

In many practical learning scenarios, there is a small amount of labeled data along with a large pool of unlabeled data. Many supervised learning algorithms have been developed and extensively studied. We present a new "co-training" strategy for using unlabeled data to improve the performance of standard supervised learning algorithms. Unlike much of the prior work, such as the co-training procedure of Blum and Mitchell (1998), we do not assume there are two redundant views both of which are sufficient for perfect classification. The only requirement our co-training strategy places on each supervised learning algorithm is that its hypothesis partitions the example space into a set of equivalence classes (e.g. for a decision tree each leaf defines an equivalence class). We evaluate our co-training strategy via experiments using data from the UCI repository. 1. Introduction In many practical learning scenarios, there is a small amount of labeled data along with a lar...

Discriminative learning methods are widely used in natural language processing. These methods work best when their training and test data are drawn from the same distribution. For many NLP tasks, however, we are confronted with new domains in which labeled data is scarce or non-existent. In such cases, we seek to adapt existing models from a resourcerich source domain to a resource-poor target domain. We introduce structural correspondence learning to automatically induce correspondences among features from different domains. We test our technique on part of speech tagging and show performance gains for varying amounts of source and target training data, as well as improvements in target domain parsing accuracy using our improved tagger. 1

We state and analyze the first active learning algorithm which works in the presence of arbitrary forms of noise. The algorithm, A2 (for Agnostic Active), relies only upon the assumption that the samples are drawn i.i.d. from a fixed distribution. We show that A2 achieves an exponential improvement (i.e., requires only O � ln 1 ɛ samples to find an ɛ-optimal classifier) over the usual sample complexity of supervised learning, for several settings considered before in the realizable case. These include learning threshold classifiers and learning homogeneous linear separators with respect to an input distribution which is uniform over the unit sphere. 1.

Co-Training is a weakly supervised learning paradigm in which the redundancy of the learning task is captured by training two classifiers using separate views of the same data. This enables bootstrapping from a small set of labeled training data via a large set of unlabeled data. This study examines the learning behavior of co-training on natural language processing tasks that typically require large numbers of training instances to achieve usable performance levels. Using base noun phrase bracketing as a case study, we find that co-training reduces by 36% the di#erence in error between co-trained classifiers and fully supervised classifiers trained on a labeled version of all available data. However, degradation in the quality of the bootstrapped data arises as an obstacle to further improvement. To address this, we propose a moderately supervised variant of cotraining in which a human corrects the mistakes made during automatic labeling. Our analysis suggests that corrected co-training and similar moderately supervised methods may help cotraining scale to large natural language learning tasks. 1

In a multi-view problem, the features of the domain can be partitioned into disjoint subsets (views) that are sufficient to learn the target concept.

The structure of the web is increasingly being used to improve organization, search, and analysis of information on the web. For example, Google uses the text in citing documents (documents that link to the target document) for search. We analyze the relative utility of document text, and the text in citing documents near the citation, for classification and description. Results show that the text in citing documents, when available, often has greater discriminative and descriptive power than the text in the target document itself. The combination of evidence from a document and citing documents can improve on either information source alone. Moreover, by ranking words and phrases in the citing documents according to expected entropy loss, we are able to accurately name clusters of web pages, even with very few positive examples. Our results confirm, quantify, and extend previous research using web sn'ucture in these areas, introducing new methods for classification and description of pages.

The approach of learning of multiple "related" tasks simultaneously has proven quite successful in practice; however, theoretical justification for this success has remained elusive. The starting point of previous work on multiple task learning has been that the tasks to be learnt jointly are somehow "algorithmically related", in the sense that the results of applying a specific learning algorithm to these tasks are assumed to be similar. We take a logical step backwards and offer a data generating mechanism through which our notion of task-relatedness is defined.