For one-sided matching problems, two widely studied mechanisms are the Random Serial Dictatorship (RSD) and the Probabilistic Serial Rule (PS). The induced outcomes of these two mechanisms are often incomparable and thus there are challenges when it comes to deciding which mechanism to adopt in practice. Working in the space of general preferences, we provide empirical results on the (in)compar...

This paper describes the methods we have employed to solve the author profiling task at PAN-2013. Our goal was to use simple features to identify the age group and the gender of the author of a given text. We introduce the features, detail how the classifiers were trained, and how the experiments were run.

A robust image similarity measure is presented for patterned triangles. A novel feature of the new similarity measure is the reduction of the patterns to a simple set of affine-invariant bar-codes.

1. INTRODUCTION II [1] two of us iivestigated the problem of determiiiig those cardiials a, /3, y, 8, A for which the followiig statemeet, abbreviated Iá(8, g, a, A, y), holds : "Wheeever V is aa a-dimeesiooal vector space over a field of A elemeets, aad the 8-dimeesiooal subspaces of V are partitiooed iito y classes, there is some /-dimeesiooal subspace of V all of whose 8-dimee-siooal subspac...

The sliding three-tube (S3T) model, based on Fant’s acoustic theory and proposed in our previous studies, has a simple structure, enabling it to produce human-like vowels useful for education in acoustics and speech science. In this study, we changed the size of the S3T model and combined it with sound sources with different fundamental frequencies. We confirmed that the models could produce vo...

We describe how the use o1' at hierarchy of lexical categories instead of a simple set of categories leads to the definition of a flexible and precise language for the description of dependency structures. After specifying the t'ormalism we use to decorate these structures, we present an application aiming to detect and correct en'ors in a written text. We outline how tile use of the h ierarchy...

CLaC-CORE, an exhaustive feature combination system ranked 4th among 34 teams in the Semantic Textual Similarity shared task STS 2013. Using a core set of 11 lexical features of the most basic kind, it uses a support vector regressor which uses a combination of these lexical features to train a model for predicting similarity between sentences in a two phase method, which in turn uses all combi...