In this paper we present some lessons learned from building vizsla, the keyword search and topic classification system used on the largest Hungarian portal, [origo.hu]. Based on a simple statistical language model, and the largescale supporting evidence from vizsla, we argue that in topic classification only positive ev-

George Mackey devoted his career to exploring the basic unity of mathematics, and to understanding its relationships to modern physics. He was a major figure in the flowering of group representation theory and related portions of functional analysis that occured in the middle decades of the last century. Sadly, he, along with many of his colleagues who participated in this endeavor, is no longe...

Analysis and design patterns are well established as a convenient and reusable way to build high-quality object-oriented software. Patterns combine experience and good practices to develop basic models that can be used for new designs. Security patterns join the extensive knowledge accumulated about security with the structure provided by patterns to provide guidelines for secure system design ...

Classifying Discourse Relations Mridhula Raghupathy & Hena Mehta [email protected] | [email protected] Faculty Advisors: Dr. Aravind Joshi, Dr. Ani Nenkova, & Dr. Alan Lee Abstract The goal of this project was to study properties of discourse relations as they appear in the Penn Discourse Tree Bank (PDTB), a large corpus of naturally occurring text whose discourse relations and their fe...

Variability modeling is important for managing variability in software product families, especially during product derivation. In the past few years, several variability modeling techniques have been developed, each using its own concepts to model the variability provided by a product family. The publications regarding these techniques were written from diVerent viewpoints, use diVerent example...

This paper presents a fast algorithm for computing local symmetry descriptions of region shape. Like previous algorithms, it extracts groups of edge points tangent to a common circle. However, by examining the number of distinct points of tangency, the new algorithm separates groups belonging to round regions from those belonging to elongated regions. Thus, when these relations are connected to...

We prove a consistency results saying, that for a simple (first order) theory, it is easier to have a universal model in some cardinalities, than for the theory of linear order. We define additional properties of first order theories, the n-strong order property (SOPn in short). The main result is that a first order theory with the 4-strong order property behaves like linear orders concerning e...

Let {Σα}α<ω1 , where ω1 is the first non-countable ordinal, denote the Borel hierarchy of subsets of the Cantor space 2. As usual, Πα denotes the dual class for Σα while ∆ 0 α = Σ 0 α ∩Πα — the corresponding ambiguous class. Let B = ∪α<ω1Σα denote the class of all Borel sets. The levels of Borel hierarchy, as well as many other classes of interest for descriptive set theory, may be defined by m...

The present paper studies a particular collection of classi cation problems, i.e., the classi cation of recursive predicates and languages, for arriving at a deeper understanding of what classi cation really is. In particular, the classi cation of predicates and languages is compared with the classi cation of arbitrary recursive functions and with their learnability. The investigation undertake...