Base de données orientée graphe

Une base de données orientée graphe est une base de données orientée objet utilisant la théorie des graphes, donc avec des nœuds et des arcs, permettant de représenter et stocker les données. Qu'est ce qu'une base de données orientée graphe ? Par définition, une base de données orientée graphe correspond à un système de stockage capable de fournir une adjacence entre éléments voisins : chaque voisin d'une entité est accessible grâce à un pointeur physique. C'est une base de données orientée objet adaptée à l'exploitation des structures de données de type graphe ou dérivée, comme des arbres. Exemple d'illustration : En base de données relationnelle traditionnelle : La table Client contient les champs suivants : un identifiant, un nom, un prénom, une adresse, un numéro client. Chacun de ces noms de champs correspond aux noms des colonnes respectives. Considérons l'enregistrement d'un client qui a les valeurs suivantes comme couples "(champs=valeur)" : (identifiant=id

Filtering Random Graph Processes Over Random Time-Varying Graphs

Andreas Loukas

Graph filters play a key role in processing the graph spectra of signals supported on the vertices of a graph. However, despite their widespread use, graph filters have been analyzed only in the deterministic setting, ignoring the impact of stochasticity in both the graph topology and the signal itself. To bridge this gap, we examine the statistical behavior of the two key filter types, finite impulse response and autoregressive moving average graph filters, when operating on random time-varying graph signals (or random graph processes) over random time-varying graphs. Our analysis shows that 1) in expectation, the filters behave as the same deterministic filters operating on a deterministic graph, being the expected graph, having as input signal a deterministic signal, being the expected signal, and 2) there are meaningful upper bounds for the variance of the filter output. We conclude this paper by proposing two novel ways of exploiting randomness to improve (joint graph-time) noise cancellation, as well as to reduce the computational complexity of graph filtering. As demonstrated by numerical results, these methods outperform the disjoint average and denoise algorithm and yield a (up to) four times complexity reduction, with a very little difference from the optimal solution.

Institute of Electrical and Electronics Engineers2017

Scalable Robust Graph Embedding with Spark

Karl Aberer, Chi Thang Duong, Trung-Dung Hoang, Quoc Viet Hung Nguyen

Graph embedding aims at learning a vector-based representation of vertices that incorporates the structure of the graph. This representation then enables inference of graph properties. Existing graph embedding techniques, however, do not scale well to large graphs. While several techniques to scale graph embedding using compute clusters have been proposed, they require continuous communication between the compute nodes and cannot handle node failure. We therefore propose a framework for scalable and robust graph embedding based on the MapReduce model, which can distribute any existing embedding technique. Our method splits a graph into subgraphs to learn their embeddings in isolation and subsequently reconciles the embedding spaces derived for the subgraphs. We realize this idea through a novel distributed graph decomposition algorithm. In addition, we show how to implement our framework in Spark to enable efficient learning of effective embeddings. Experimental results illustrate that our approach scales well, while largely maintaining the embedding quality.

ASSOC COMPUTING MACHINERY2021

Biomechanics at the workplace under hypogravity conditions

Tatiana Volkova

The main goal of my research is to establish guidelines for workplace design based on human biomechanics: specifically sitting workplaces and handling areas in 1/6G-1/3G (Moon, Mars) conditions. Such a workplace could be used in long-term space missions in order to maximize worker performance and minimize the risks of musculoskeletal injuries.Research objective is to formalize a baseline workplace design based on best practices limited to the tasks: operating a computer, assembling/maintenance operations.Proposed steps:1) Data collection and methods definitions2) Measurement and simulation of the human fatigue in the variation of sitting postures and handling areas (identification of typical poses, 3D simulation, graph analysis)3) Optimization of the workplace design based on human biomechanics (motion capture, factor analysis, optimization) 4) Simulation of workplace design in different environments (validation methods).Scientific & technological novelty: 1) Measurement of human muscular fatigue as a function of workload in reduced gravity conditions and integration of the results into the design model2) Modeling of the biomechanics of human movements in the workplace in reduced gravity conditions;3) Developing of sitting workplace design in reduced gravity conditions - taking into account human biomechanics.This research addresses problems revealed by workplace design in confined conditions found in Earth or near-Earth facilities in extreme environments including in underwater conditions, remote & non-remote locations and Low Earth Orbit (LEO), habitat simulators. Human-centered design strategies and biomechanics issues for creating an effective workplace design for long-term missions will also be addressed. The recommendations for design, maintenance, and usage of this workplace in different gravity conditions (1G, 1/6G,1/3G) will be provided. The design optimization of such a workplaces design based on human biomechanics will be done with the help of numerical simulations and digital human modeling. Physical experiments will validate these results by means of the suspension system, underwater experiment and possible by parabolic flights. This work may have a significant impact on workplace designs for both Earth and Space projects. Keywords: biomechanics, workplace, posture, extreme environments, human-centered design, Moon/Mars exploration, comfort conditions, task performance, reduced gravity, digital human modeling, motion capture

EPFL2022

Biomechanics at the workplace under hypogravity conditions

Tatiana Volkova

EPFL2022