Martha Sideri

BlogRank: ranking weblogs based on connectivity and similarity features

A large part of the hidden web resides in weblog servers; traditional search engines perform poorly on blogs. We present a method for ranking weblogs utilizing both link graph and similarity, and based on an enhanced and weighted graph of weblogs capturing crucial weblog features. Rankings are then assigned using our algorithm, BlogRank, which is a modified version of PageRank.To validate our method we ran experiments on a weblog dataset, processed and adapted to our search engine: http://spiderwave.aueb.gr/BlogwaveOur experiments suggest that our algorithm enhances the quality of returned results.

On the complexity of reconfiguration problems

Reconfiguration problems arise when we wish to find a step-by-step transformation between two feasible solutions of a problem such that all intermediate results are also feasible. We demonstrate that a host of reconfiguration problems derived from NP-complete problems are PSPACE-complete, while some are also NP-hard to approximate. In contrast, several reconfiguration versions of problems in P are solvable in polynomial time.

On Horn Envelopes and Hypergraph Transversals

We study the problem of bounding from above and below a given set of bit vectors by the set of satisfying truth assignments of a Horn formula. We point out a rather unexpected connection between the upper bounding problem and the problem of generating all transversals of a hypergraph, and settle several related complexity questions.

Fixed-parameter complexity in AI and nonmonotonic reasoning

Many relevant intractable problems become tractable if some problem parameter is fixed. However, various problems exhibit very different computational properties, depending on how the runtime required for solving them is related to the fixed parameter chosen. The theory of parameterized complexity deals with such issues, and provides general techniques for identifying fixed-parameter tractable and fixed-parameter intractable problems. We study the parameterized complexity of various problems in AI and nonmonotonic reasoning. We show that a number of relevant parameterized problems in these areas are fixed-parameter tractable. Among these problems are constraint satisfaction problems with bounded treewidth and fixed domain, restricted forms of conjunctive database queries, restricted satisfiability problems, propositional logic programming under the stable model semantics where the parameter is the dimension of a feedback vertex set of the programs dependency graph, and circumscriptive inference from a positive k-CNF restricted to models of bounded size. We also show that circumscriptive inference from a general propositional theory, when the attention is restricted to models of bounded size, is fixed-parameter intractable and is actually complete for a novel fixed-parameter complexity class.

Generating all maximal models of a Boolean expression

We examine the computational problem of generating all maximal models of a Boolean expression in CNF. We give a resolution-like method that reduces the unnegated variables of an expression while preserving its set of maximal models. We present an output-polynomial algorithm for the 2CNF case and we show that the problem cannot be solved in outputpolynomial time in the case of Horn expressions, unless PD NP, despite an affinity of this case to the recently subexponentially solved transversal hypergraph problem. The problem is of course trivial for 1-valid and anti-Horn expressions, and open for exclusive-ors; it is NP-hard in all other cases.

The compass filter: search engine result personalization using web communities

We propose a simple approach to search engine personalization based on Web communities [14]. User information –in particular, the Web communities whose neighborhoods the user has selected in the past– is used to change the order of the returned search results. We present experimental evidence suggesting that our method indeed improves the quality of the ranking. Our experiments were carried out on a search engine created by our research group and focusing on the Greek fragment of the worldwide Web (1.33 million documents); we also discuss the issue of scaling.

The bisection width of grid graphs

A solid grid graph is a finite induced subgraph of the infinite grid that has no holes. We present a polynomial algorithm for computing the minimum number of edges we need to delete in order to divide a given solid grid graph into two parts containing an equal number of nodes. The algorithm is based on dynamic programming, and it extends to several related problems, including grid graphs with a bounded number of holes.

Default theories that always have extensions

Default logic is an important and influential formalization of commonsense reasoning. Determining whether a given default theory has an extension is the main computational problem pertinent to default logic, the analog of testing for validity and finding deductions in more classical logics. Substantially generalizing a result by Etherington [5] we show that all default theories that have no odd cycles (in some precise sense) have an extension, which can be found efficiently. We also give a proof that it is NP-complete to find extensions even for default theories with no prerequisites and at most two literals per default, a case substantially simpler than the NP-completeness results in the literature.

Fixed-Parameter Complexity in AI and Nonmonotonic Reasoning

We study the fixed-parameter complexity of various problems in AI and nonmonotonic reasoning. We show that a number of relevant parameterized problems in these areas are fixed-parameter tractable. Among these problems are constraint satisfaction problems with bounded treewidth and fixed domain, restricted satisfiability problems, propositional logic programming under the stable model semantics where the parameter is the dimension of a feedback vertex set of the programs dependency graph, and circumscriptive inference from a positive k-CNF restricted to models of bounded size. We also show that circumscriptive inference from a general propositional theory, when the attention is restricted to models of bounded size, is fixed-parameter intractable and is actually complete for a novel fixed-parameter complexity class.

On the Floyd-Warshall Algorithm for Logic Programs

We explore the possibility of evaluating single-rule Datalog programs efficiently and with logarithmic work space by a natural extension of the Floyd–Warshall algorithm for transitive closure. We characterize exactly the single rule chain programs that can be so evaluated – they are rather modest generalizations of the transitive closure. The proof relies on an interesting language-theoretic concept, total ambiguity. Extensions to more general classes of programs, and more general algorithms, are discussed.