We present a family of margin based online learning algorithms for various prediction tasks. In particular we derive and analyze algorithms for binary and multiclass categorization, regression, uniclass prediction and sequence prediction. The update steps of our different algorithms are all based on analytical solutions to simple constrained optimization problems. This unified view allows us to prove worst-case loss bounds for the different algorithms and for the various decision problems based on a single lemma. Our bounds on the cumulative loss of the algorithms are relative to the smallest loss that can be attained by any fixed hypothesis, and as such are applicable to both realizable and unrealizable settings. We demonstrate some of the merits of the proposed algorithms in a series of experiments with synthetic and real data sets.

We study the problem of classifying data in a given taxonomy when classifications associated with multiple and/or partial paths are allowed. We introduce a new algorithm that incrementally learns a linear-threshold classifier for each node of the taxonomy. A hierarchical classification is obtained by evaluating the trained node classifiers in a top-down fashion. To evaluate classifiers in our multipath framework, we define a new hierarchical loss function, the H-loss, capturing the intuition that whenever a classification mistake is made on a node of the taxonomy, then no loss should be charged for any additional mistake occurring in the subtree of that node. Making no assumptions on the mechanism generating the data instances, and assuming a linear noise model for the labels, we bound the H-loss of our on-line algorithm in terms of the H-loss of a reference classifier knowing the true parameters of the label-generating process. We show that, in expectation, the excess cumulative H-loss grows at most logarithmically in the length of the data sequence. Furthermore, our analysis reveals the precise dependence of the rate of convergence on the eigenstructure of the data each node observes. Our theoretical results are complemented by a number of experiments on texual corpora. In these experiments we show that, after only one epoch of training, our algorithm performs much better than Perceptron-based hierarchical classifiers, and reasonably close to a hierarchical support vector machine.

This paper suggests a method for multiclass learning with many classes by simultaneously learning shared characteristics common to the classes, and predictors for the classes in terms of these characteristics. We cast this as a convex optimization problem, using trace-norm regularization and study gradient-based optimization both for the linear case and the kernelized setting. 1.

We present a kernel-based algorithm for hierarchical text classification where the documents are allowed to belong to more than one category at a time. The classification model is a variant of the Maximum Margin Markov Network framework, where the classification hierarchy is represented as a Markov tree equipped with an exponential family defined on the edges. We present an efficient optimization algorithm based on incremental conditional gradient ascent in single-example subspaces spanned by the marginal dual variables. The optimization is facilitated with a dynamic programming based algorithm that computes best update directions in the feasible set. Experiments show

We describe a general framework for online multiclass learning based on the notion of hypothesis sharing. In our framework sets of classes are associated with hypotheses. Thus, all classes within a given set share the same hypothesis. This framework includes as special cases commonly used constructions for multiclass categorization such as allocating a unique hypothesis for each class and allocating a single common hypothesis for all classes. We generalize the multiclass Perceptron to our framework and derive a unifying mistake bound analysis. Our construction naturally extends to settings where the number of classes is not known in advance but, rather, is revealed along the online learning process. We demonstrate the merits of our approach by comparing it to previous methods on both synthetic and natural datasets. 1.

We study hierarchical classification in the general case when an instance could belong to more than one class node in the underlying taxonomy. Experiments done in previous work showed that a simple hierarchy of Support Vectors Machines (SVM) with a top-down evaluation scheme has a surprisingly good performance on this kind of task. In this paper, we introduce a refined evaluation scheme which turns the hierarchical SVM classifier into an approximator of the Bayes optimal classifier with respect to a simple stochastic model for the labels. Experiments on synthetic datasets, generated according to this stochastic model, show that our refined algorithm outperforms the simple hierarchical SVM. On real-world data, however, the advantage brought by our approach is a bit less clear. We conjecture this is due to a higher noise rate for the training labels in the low levels of the taxonomy. 1.

While large-scale taxonomies – especially for web pages – have been in existence for some time, approaches to automatically classify documents into these taxonomies have met with limited success compared to the more general progress made in text classification. We argue that this stems from three causes: increasing sparsity of training data at deeper nodes in the taxonomy, error propagation where a mistake made high in the hierarchy cannot be recovered, and increasingly complex decision surfaces in higher nodes in the hierarchy. While prior research has focused on the first problem, we introduce methods that target the latter two problems – first by biasing the training distribution to reduce error propagation and second by propagating up “first-guess ” expert information in a bottom-up manner before making a refined top down choice. Finally, we present an empirical study demonstrating that the suggested changes lead to 10-30 % improvements in F1 scores versus an accepted competitive baseline, hierarchical SVMs.

Recent work on ontology-based Information Extraction (IE) has tried to make use of knowledge from the target ontology in order to improve semantic annotation results. However, very few approaches exploit the ontology structure itself, and those that do so, have some limitations. This paper introduces a hierarchical learning approach for IE, which uses the target ontology as an essential part of the extraction process, by taking into account the relations between concepts. The approach is evaluated on the largest available semantically annotated corpus. The results demonstrate clearly the benefits of using knowledge from the ontology as input to the information extraction process. We also demonstrate the advantages of our approach over other state-of-the-art learning systems on a commonly used benchmark dataset.

Many real-world classification problems involve large numbers of overlapping categories that are arranged in a hierarchy or taxonomy. We propose to incorporate prior knowledge on category taxonomy directly into the learning architecture. We present two concrete multi-label classification methods, a generalized version of Perceptron and a hierarchical multi-label SVM learning. Our method works with arbitrary, not necessarily singly connected taxonomies, and can be applied more generally in settings where categories are characterized by attributes and relations that are not necessarily induced by a taxonomy. Experimental results on WIPO-alpha collection show that our hierarchical methods bring significant performance improvement. 1

Prediction suffix trees (PST) provide a popular and effective tool for tasks such as compression, classification, and language modeling. In this paper we take a decision theoretic view of PSTs for the task of sequence prediction. Generalizing the notion of margin to PSTs, we present an online PST learning algorithm and derive a loss bound for it. The depth of the PST generated by this algorithm scales linearly with the length of the input. We then describe a self-bounded enhancement of our learning algorithm which automatically grows a bounded-depth PST. We also prove an analogous mistake-bound for the self-bounded algorithm. The result is an efficient algorithm that neither relies on a-priori assumptions on the shape or maximal depth of the target PST nor does it require any parameters. To our knowledge, this is the first provably-correct PST learning algorithm which generates a bounded-depth PST while being competitive with any fixed PST determined in hindsight. 1