Indexing and querying segmented web pages: the BlockWeb Model

We present in this paper a model for indexing and querying web pages, based on the hierarchical decomposition of pages into blocks. Splitting up a page into blocks has several advantages in terms of page design, indexing and querying such as (i) blocks of a page most similar to a query may be returned instead of the page as a whole (ii) the importance of a block can be taken into account, as well as (iii) the permeability of the blocks to neighbor blocks: a block b is said to be permeable to a block b?? in the same page if b?? content (text, image, etc.) can be (partially) inherited by b upon indexing. An engine implementing this model is described including: the transformation of web pages into blocks hierarchies, the definition of a dedicated language to express indexing rules and the storage of indexed blocks into an XML repository. The model is assessed on a dataset of electronic news, and a dataset drawn from web pages of the ImagEval campaign where it improves by 16% the mean average precision of the baseline.     

Learning a generic 3D face model from 2D image databases using incremental Structure-from-Motion

Over the last decade 3D face models have been extensively used in many applications such as face recognition, facial animation and facial expression analysis. 3D Morphable Models (MMs) have become a popular tool to build and fit 3D face models to images. Critical to the success of MMs is the ability to build a generic 3D face model. Major limitations in the MMs building process are: (1) collecting 3D data usually involves the use of expensive laser scans and complex capture setups, (2) the number of available 3D databases is limited, and typically there is a lack of expression variability and (3) finding correspondences and registering the 3D model is a labor intensive and error prone process.     

A tabu search with an oscillation strategy for the discriminant analysis problem

This article proposes a tabu search approach to solve a mathematical programming formulation of the linear classification problem, which consists of determining an hyperplane that separates two groups of points as well as possible in ?^m. The tabu search approach proposed is based on a non-standard formulation using linear system infeasibility. The search space is the set of bases defined on the matrix that describes the linear system. The moves are performed by pivoting on a specified row and column. On real machine learning databases, our approach compares favorably with implementations based on parametric programming and irreducible infeasible constraint sets. Additional computational results for randomly generated instances confirm that our method provides a suitable alternative to the mixed integer programming formulation that is solved by a commercial code when the number of attributes m increases.     

Connecting the 3D DGS Calques3D with the CAS Maple

Many (2D) Dynamic Geometry Systems (DGSs) are able to export numeric coordinates and equations with numeric coefficients to Computer Algebra Systems (CASs). Moreover, different approaches and systems that link (2D) DGSs with CASs, so that symbolic coordinates and equations with symbolic coefficients can be exported from the DGS to the CAS, already exist. Although the 3D DGS Calques3D can export numeric coordinates and equations with numeric coefficients to Maple and Mathematica, it cannot export symbolic coordinates and equations with symbolic coefficients. A connection between the 3D DGS Calques3D and the CAS Maple, that can handle symbolic coordinates and equations with symbolic coefficients, is presented here. Its main interest is to provide a convenient time-saving way to explore problems and directly obtain both algebraic and numeric data when dealing with a 3D extension of “ruler and compass geometry”. This link has not only educational purposes but mathematical ones, like mechanical theorem proving in geometry, geometric discovery (hypotheses completion), geometric loci finding… As far as we know, there is no comparable “symbolic” link in the 3D case, except the prototype 3D-LD (restricted to determining algebraic surfaces as geometric loci).     

A branch-and-cut algorithm for the minimum labeling Hamiltonian cycle problem and two variants

This paper proposes a mathematical model, valid inequalities and polyhedral results for the minimum labeling Hamiltonian cycle problem. This problem is defined on an unweighted graph in which each edge has a label. The aim is to determine a Hamiltonian cycle with the least number of labels. We also define two variants of this problem by assigning weights to the edges and by considering the tour length either as an objective or as a constraint. A branch-and-cut algorithm for the three problems is developed, and computational results are reported on randomly generated instances and on modified instances from TSPLIB.     

Nondeterministic Graph Searching: From Pathwidth to Treewidth

We introduce nondeterministic graph searching with a controlled amount of nondeterminism and show how this new tool can be used in algorithm design and combinatorial analysis applying to both pathwidth and treewidth. We prove equivalence between this game-theoretic approach and graph decompositions called q -branched tree decompositions, which can be interpreted as a parameterized version of tree decompositions. Path decomposition and (standard) tree decomposition are two extreme cases of q-branched tree decompositions. The equivalence between nondeterministic graph searching and q-branched tree decomposition enables us to design an exact (exponential time) algorithm computing q-branched treewidth for all q≥0, which is thus valid for both treewidth and pathwidth. This algorithm performs as fast as the best known exact algorithm for pathwidth. Conversely, this equivalence also enables us to design a lower bound on the amount of nondeterminism required to search a graph with the minimum number of searchers. Additional support of F.V. Fomin by the Research Council of Norway. Additional supports of P. Fraigniaud from the INRIA Project “Grand Large”, and from the Project PairAPair of the ACI “Masse de Données”. Additional supports of N. Nisse from the Project Fragile of the ACI “Sécurité & Informatique”.     

Graph searching with advice

Fraigniaud et al. [L. Blin, P. Fraigniaud, N. Nisse, S. Vial, Distributing chasing of network intruders, in: 13th Colloquium on Structural Information and Communication Complexity, SIROCCO, in: LNCS, vol. 4056, Springer-Verlag, 2006, pp. 70–84] introduced a new measure of difficulty for a distributed task in a network. The smallest number of bits of advice   of a distributed problem is the smallest number of bits of information that has to be available to nodes in order to accomplish the task efficiently. Our paper deals with the number of bits of advice required to perform efficiently the graph searching problem in a distributed setting. In this variant of the problem, all searchers are initially placed at a particular node of the network. The aim of the team of searchers is to clear a contaminated graph in a monotone connected way, i.e., the cleared part of the graph is permanently connected, and never decreases while the search strategy is executed. Moreover, the clearing of the graph must be performed using the optimal number of searchers, i.e. the minimum number of searchers sufficient to clear the graph in a monotone connected way in a centralized setting. We show that the minimum number of bits of advice permitting the monotone connected and optimal clearing of a network in a distributed setting is Θ(nlogn)$\Theta (nlogn)$, where n$n$ is the number of nodes of the network. More precisely, we first provide a labelling of the vertices of any graph G$G$, using a total of O(nlogn)$O(nlogn)$ bits, and a protocol using this labelling that enables the optimal number of searchers to clear G$G$ in a monotone connected distributed way. Then, we show that this number of bits of advice is optimal: any distributed protocol requires Ω(nlogn)$\Omega (nlogn)$ bits of advice to clear a network in a monotone connected way, using an optimal number of searchers.     

Parallel Lattice Boltzmann Method with Blocked Partitioning

This paper presents and discusses a blocked parallel implementation of bi- and three-dimensional versions of the Lattice Boltzmann Method. This method is used to represent and simulate fluid flows following a mesoscopic approach. Most traditional parallel implementations use simple data distribution strategies to parallelize the operations on the regular fluid data set. However, it is well known that block partitioning is usually better. Such a parallel implementation is discussed and its communication cost is established. Fluid flows simulations crossing a cavity have also been used as a real-world case study to evaluate our implementation. The presented results with our blocked implementation achieve a performance up to 31% better than non-blocked versions, for some data distributions. Thus, this work shows that blocked, parallel implementations can be efficiently used to reduce the parallel execution time of the method.