We propose a PAC-Bayes sample compression approach to kernel methods that can accommodate any bounded similarity function and show that the support vector machine (SVM) classifier is a particular case of a more general class of data-dependent classifiers known as majority votes of samplecompressed classifiers. We provide novel risk bounds for these majority votes and learning algorithms that mi...

This paper describes aueb’s participation in tac 2008. Specifically, we participated in the summarization and textual entailment recognition tracks. For the former we trained a Support Vector Regression model that is used to rank the summary’s candidate sentences; and for the latter we used a Maximum Entropy classifier along with string similarity measures applied to several abstractions of the...

Our submission to the Semeval 2010 task on coreference resolution in multiple languages is based on parse analysis and similarity clustering. The system uses a binary classifier, based on Maximum Entropy, to decide whether or not there is a relationship between each pair of mentions extracted from a textual document. Mention detection is based on the analysis of the dependency parse tree.

The development of a sentiment classifier experiences two problems to cope with: the demand of large amounts of labelled training data and a decrease in performance when the classifier is applied to a different domain. In this paper, we attempt to address this problem by exploring a number of metrics that try to predict the cross-domain performance of a sentiment classifier through the analysis...

Recent results in computer vision have supported the theory that object detectors built in the statistical learning framework can benefit from a two stage learning process, first learning appropriate diagnostic features for the object being trained, and subsequently training an upper-level classifier on the excitation of these part detectors. In this study we develop a hierarchical detection ar...

A new approach, based on the k-Nearest Neighbor (kNN) classifier, is used to classify program behavior as normal or intrusive. Short sequences of system calls have been used by others to characterize a program’s normal behavior before. However, separate databases of short system call sequences have to be built for different programs, and learning program profiles involves time-consuming trainin...

In this paper, we consider the problem of feature selection and classifier fusion and discuss how they should be reflected in the fusion system architecture. We employed the genetic algorithm with a novel coding to search the worst performing fusion strategy. The proposed algorithm tunes itself between feature and matching score levels, and improves the final performance over the original on tw...

This work reports baseline results for the CLEF 2008 Medical Automatic Annotation Task (MAAT) by applying a classifier with a fixed parameter set to all tasks 2005 – 2008. A nearest-neighbor (NN) classifier is used, which uses a weighted combination of three distance and similarity measures operating on global image features: Scaled-down representations of the images are compared using models f...

A thesaurus is a reference work that lists words grouped together according to similarity of meaning (containing synonyms and sometimes antonyms), in contrast to a dictionary, which contains definitions and pronunciations. This paper proposes an innovative approach to improve the classification performance of Persian texts considering a very large thesaurus. The paper proposes a flexible method...

The immune system is capable of learning, memory, and pattern recognition. By employing genetic operators on a time scale fast enough to observe experimentally, the immune system is able to recognize novel shapes without preprogramming. Here we describe a dynamical model for the immune system that is based on the network hypothesis of Jerne, and is simple enough to simulate on a computer. This ...