Auswahl der wissenschaftlichen Literatur zum Thema „Système de k-graphes“

Since the discovery of the quantum Hall effect in 1980, it has attracted intense interest in condensed matter physics and has led to a new type of metrological standard by utilizing the resistance quantum. Graphene, a true two-dimensional electron gas material, has demonstrated the half-integer quantum Hall effect and composite-fermion fractional quantum Hall effect due to its unique massless Dirac fermions and ultra-high carrier mobility. Here, we use a monolayer graphene encapsulated with hexagonal boron nitride and few-layer graphite to fabricate micrometer-scale graphene Hall devices. The application of a graphite gate electrode significantly screens the phonon scattering from a conventional SiO2/Si substrate, and thus enhances the carrier mobility of graphene. At a low temperature, the carrier mobility of graphene devices can reach 3 × 105 cm2/V·s, and at room temperature, the carrier mobility can still exceed 1 × 105 cm2/V·s, which is very helpful for the development of high-temperature quantum Hall effects under moderate magnetic fields. At a low temperature of 1.6 K, a series of half-integer quantum Hall plateaus are well-observed in graphene with a magnetic field of 1 T. More importantly, the ν = ±2 quantum Hall plateau clearly persists up to 150 K with only a few-tesla magnetic field. These findings show that graphite-gated high-mobility graphene devices hold great potential for high-sensitivity Hall sensors and resistance metrology standards for the new Système International d’unités.

Graphene membranes act as temperature sensors in nanoelectromechanical devices due to their excellent thermal and high-temperature resistance properties. Experimentally, reports on the sensing performance of graphene mainly focus on the temperature interval under 400 K. To explore the sensing performance of graphene temperature sensors at higher temperature intervals, micro-fabricated single-layer graphene on a SiNX substrate is presented as temperature sensors by semiconductor technology and its electrical properties were measured. The results show that the temperature coefficient of the resistance value is 2.07 × 10−3 in the temperature range of 300–450 K and 2.39 × 10−3 in the temperature range of 450–575 K. From room temperature to high temperature, the “metal” characteristics are presented, and the higher TCR obtained at higher temperature interval is described and analyzed by combining Boltzmann transport equation and thermal expansion theory. These investigations provide further insight into the temperature characteristics of graphene.

As an effective technique for fabricating conductive and thermally conductive polymer composites, a multi-filler system incorporates different types and sizes of multiple fillers to form interconnected networks with improved electrical, thermal, and processing properties. In this study, DIW forming of bifunctional composites was achieved by controlling the temperature of the printing platform. The study was based on enhancing the thermal and electrical transport properties of hybrid ternary polymer nanocomposites with multi-walled carbon nanotubes (MWCNTs) and graphene nanoplates (GNPs). With thermoplastic polyurethane (TPU) used as the matrix, the addition of MWCNTs, GNPs and both mixtures further improved the thermal conductivity of the elastomers. By adjusting the weight fraction of the functional fillers (MWCNTs and GNPs), the thermal and electrical properties were gradually explored. Here, the thermal conductivity of the polymer composites increased nearly sevenfold (from 0.36 W·m−1·k−1 to 2.87 W·m−1·k−1) and the electrical conductivity increased up to 5.49 × 10−2 S·m−1. It is expected to be used in the field of electronic packaging and environmental thermal dissipation, especially for modern electronic industrial equipment.

Resistive random access memory (RRAM), which is considered as one of the most promising next-generation non-volatile memory (NVM) devices and a representative of memristor technologies, demonstrated great potential in acting as an artificial synapse in the industry of neuromorphic systems and artificial intelligence (AI), due its advantages such as fast operation speed, low power consumption, and high device density. Graphene and related materials (GRMs), especially graphene oxide (GO), acting as active materials for RRAM devices, are considered as a promising alternative to other materials including metal oxides and perovskite materials. Herein, an overview of GRM-based RRAM devices is provided, with discussion about the properties of GRMs, main operation mechanisms for resistive switching (RS) behavior, figure of merit (FoM) summary, and prospect extension of GRM-based RRAM devices. With excellent physical and chemical advantages like intrinsic Young’s modulus (1.0 TPa), good tensile strength (130 GPa), excellent carrier mobility (2.0 × 105 cm2∙V−1∙s−1), and high thermal (5000 Wm−1∙K−1) and superior electrical conductivity (1.0 × 106 S∙m−1), GRMs can act as electrodes and resistive switching media in RRAM devices. In addition, the GRM-based interface between electrode and dielectric can have an effect on atomic diffusion limitation in dielectric and surface effect suppression. Immense amounts of concrete research indicate that GRMs might play a significant role in promoting the large-scale commercialization possibility of RRAM devices.

Machine learning and computer vision algorithms can provide a precise and automated interpretation of medical videos. The segmentation of the left ventricle of echocardiography videos plays an essential role in cardiology for carrying out clinical cardiac diagnosis and monitoring the patient’s condition. Most of the developed deep learning algorithms for video segmentation require an enormous amount of labeled data to generate accurate results. Thus, there is a need to develop new semi-supervised segmentation methods due to the scarcity and costly labeled data. In recent research, semi-supervised learning approaches based on graph signal processing emerged in computer vision due to their ability to avail the geometrical structure of data. Video object segmentation can be considered as a node classification problem. In this paper, we propose a new approach called GraphECV based on the use of graph signal processing for semi-supervised learning of video object segmentation applied for the segmentation of the left ventricle in echordiography videos. GraphECV includes instance segmentation, extraction of temporal, texture and statistical features to represent the nodes, construction of a graph using K-nearest neighbors, graph sampling to embed the graph with small amount of labeled nodes or graph signals, and finally a semi-supervised learning approach based on the minimization of the Sobolov norm of graph signals. The new algorithm is evaluated using two publicly available echocardiography videos, EchoNet-Dynamic and CAMUS datasets. The proposed approach outperforms other state-of-the-art methods under challenging background conditions.

Scanning electrochemical cell microscopy (SECCM) is a versatile scanning probe imaging technique that allows for simultaneous elucidation of structure activity relationships at the nanoscale in defined electrolyte volumes and provides high resolution (nm length scale) information on the topography of surfaces and interfaces1. Since its inception in 2010 SECCM has improved understanding of model systems such as graphene, graphite, carbon nanotubes, nanoparticles and conductive diamond and provided electrochemists with a tool for true single entity measurements. Electrodeposition of microscale thin films has previously been reported2 but not of crystalline species, which is the focus of this work. SECCM can also be used as a tool for the precise delivery of nano to microscale solution droplets to surfaces. By using optically transparent yet conductive substrates (such as indium tin oxide (ITO), gold or graphene ) we can couple SECCM to highly surface-sensitive imaging modes such as interference reflectance microscopy (IRM)3 or traditional brightfield microscopy, allowing the evaporation and consequent precipitation of solutes to be monitored. The SECCM apparatus allows the deposition of large arrays, under well-defined conditions (deposition time, surface potential), with each droplet representing an individual and independent crystallisation experiment and provides statistical data based on single particle level. The emergence of precipitates within these arrays is tracked to investigate the effect of additives and also surface potential on crystallisation. SECCM can use capacitative response to examine minimally ionised systems, non-aqueous solvents etc allowing organic crystals to be studied. This work focuses on the precipitation of two such model organic systems: l-cystine and 5-methyl-2-[(2-nitrophenyl)amino]thiophene-3-carbonitrile (ROY). l-Cystine is of biological relevance due to its role in the formation of kidney stone’s which are resistant to traditional therapies4. The need therefore arises for l-cystine crystallisation inhibitors making it a prime candidate for testing additive screening procedures. ROY is investigated as a model system5 to identify the merits of SECCM-IRM as a polymorph screen method. We show how sensitive precipitation is to substrate, applied potential and solvent system. References Ebejer, N.; Güell, A. G.; Lai, S. C. S.; McKelvey, K.; Snowden, M. E.; Unwin, P. R., Scanning Electrochemical Cell Microscopy: A Versatile Technique for Nanoscale Electrochemistry and Functional Imaging. Annual Review of Analytical Chemistry 2013, 6 (1), 329-351. Aaronson, B. D. B.; Garoz-Ruiz, J.; Byers, J. C.; Colina, A.; Unwin, P. R., Electrodeposition and Screening of Photoelectrochemical Activity in Conjugated Polymers Using Scanning Electrochemical Cell Microscopy. Langmuir 2015, 31 (46), 12814-12822. Valavanis, D.; Ciocci, P.; Gabriel; Morris, P.; Lemineur, J.-F.; McPherson, I. J.; Kanoufi, F.; Unwin, P. R., Hybrid scanning electrochemical cell microscopy-interference reflection microscopy (SECCM-IRM): tracking phase formation on surfaces in small volumes. Faraday Discussions 2022. Rimer, J. D.; An, Z.; Zhu, Z.; Lee, M. H.; Goldfarb, D. S.; Wesson, J. A.; Ward, M. D., Crystal Growth Inhibitors for the Prevention of l -Cystine Kidney Stones Through Molecular Design. Science 2010, 330 (6002), 337-341. Lévesque, A.; Maris, T.; Wuest, J. D., ROY Reclaims Its Crown: New Ways To Increase Polymorphic Diversity. Journal of the American Chemical Society 2020, 142 (27), 11873-11883.

Heteroatom-doped carbon materials have high gravimetric potassium-ion storage capability because of their abundant active sites and defects. However, their practical applications toward potassium storage are limited by sluggish reaction kinetics and short cycling life owing to the large ionic radius of K+ and undesirable parasitic reactions. Herein, we report a new strategy that allows for bottom-up patterning of thin N/P co-doped carbon layers with a uniform mesoporous structure on two-dimensional graphene sheets. The highly porous architecture and N/P co-doping properties provide abundant active sites for K+, and the graphene sheets promote charge/electron transfer. This synergistic structure enables excellent K+ storage performance in terms of specific capacity (387.6 mAh g-1 at 0.05 A g-1), rate capability (over 5 A g-1), and cycling stability (70% after 3,000 cycles). As a proof of concept, a potassium-ion capacitor assembled using this carbon anode yields a high energy density of 107 Wh kg-1, a maximum power density of 18.3 kW kg-1, and ultra-long cycling stability over 40,000 cycles.

La combinatoire extrémale est l'une des branches les plus vigoureuses des mathématiques combinatoires au cours des dernières décennies, et elle a été largement utilisée en informatique, en conception de réseaux et en conception de codage. Elle se concentre sur la détermination de la taille maximale ou minimale possible de certaines structures combinatoires, sous certaines conditions ou contraintes. Les ensembles hôtes peuvent être des graphes, des digraphes, des graphes aléatoires, des hypergraphes, des entiers, des nombres premiers, des ensembles, des graphes avec arêtes colorées, etc. Les structures locales peuvent être des appariements, des cliques, des cycles, des arbres, des sous-graphes couvrants (facteurs F, cycles Hamiltoniens), des familles d'intersection, des progressions arithmétiques, des solutions pour certaines équations (par exemple, x+y=z), des sous-graphes arc-en-ciel, etc. En particulier, la théorie des graphes extrémaux est une branche importante de la combinatoire extrémale, qui traite principalement de la manière dont les propriétés générales d'un graphe contrôlent la structure locale du graphe. Nous étudions l'existence d'un cycle Hamiltonien rainbow dans les systèmes de k-graphes, l'existence d'un appariement parfait rainbow dans les systèmes de k-graphes et l'existence d'un cycle long arc-en-ciel dans des graphes correctement colorés par les arêtes
Extremal Combinatorics is one of the most vigorous branch of Combinatorial Mathematics in recent decades and it has been widely used in Computer Science, Network Design and Coding Design. It focuses on determining the maximum or minimum possible size of certain combinatorial structures, subject to certain conditions or constraints. The host sets could be graphs, digraphs, random graphs, hypergraphs, integers, primes, sets, edge-colored graphs and so on. The local structures could be matchings, cliques, cycles, trees, spanning subgraphs (F-factors, Hamilton cycles), intersecting families, arithmetic progressions, solutions for some equations (e.g. x₊y₌z), rainbow subgraphs and so on. In particular, Extremal Graph Theory is a significant branch of Extremal Combinatorics, which primarily explores how the overall properties of a graph influence its local structures. We study the existence of a rainbow Hamilton cycle in k-graph systems, the existence of rainbow perfect matching in k-graph systems, and the existence of long rainbow cycle in properly edge-colored graphs

Les travaux de cette thèse se situent dans le cadre de la théorie des graphes et traitent de certains problèmes de coloration et d’autres paramètres qui leur sont associés. Les résultats obtenus trouvent des applications dans des problématiques liées aux réseaux de télécommunications. Principalement, notre contribution porte sur la résolution de problèmes et conjectures de la littérature et sur l’élargissement des connaissances dans le domaine. En premier nous présentons une étude pour différentes valeurs de k du paramètre radio kétiquetage introduit par Chartrand et al. En 2002 , qui consiste à attribuer une étiquette à chaque sommet de sorte que l’écart entre les étiquettes de sommets à distance d soit au moins k+1−d. Le but est de minimiser la différence entre la plus grande et la plus petite étiquette utilisée, appelée plage. Nous introduisons la notion de code de Gray généralisée afin d’obtenir des résultats exacts pour certains graphes. Nous démontrons également une conjecture de Chartrand et al. De 2002 sur les chaînes. Le deuxième paramètre étudié est la coloration fractionnaire totale pour des graphes circulants qui consiste à attribuer aux élément du graphe (sommets ou arêtes) des fractions de couleurs au lieu de couleurs (entiers). Le nombre chromatique fractionnaire est le ratio minimum a/b tel que l’on puisse affecter b couleurs à chaque sommet parmi un ensemble de a couleurs de sorte que deux sommets voisins aient des ensembles de couleurs disjoints. A l’aide de la définition du stable fractionnaire équilibré et complet qu’on a introduit, nous montrons un majorant du nombre chromatique fractionnaire pour les graphes circulants cubiques et des résultats exacts pour certains graphes circulants 4-réguliers
The work of this thesis placed in the context of graph theory which presents a study of some coloring problems and other parameters associated with them. The results are applied in issues related to telecommunications networks. Mainly, our contribution focuses on solving problems and conjectures in the literature and on the expansion of knowledge in the field. In first we present a study for different values of k of the radio k-labeling problem introduced by Chartrand et al. In 2002 which consists to assign a label to each vertex of a graph so that the difference between the labels of two vertices at distance d is at least k − d + 1. The goal is to minimize the difference between the largest and smallest label used, called span. We introduce the concept of generalized Gray Code in order to obtain exact results for some graphs. We solve the conjecture of Chartrand et al. 2002 about the antipodal number of the path. The second parameter studied is the fractional total coloring for circulant graphs which consist to assign a fractions of colors to each element (vertex or edge) instead of colors (integers). The fractional chromatic number is the a minimum ratio a/b such that we can assigne a sets of colors of size b to the elements of a graph from a set of a colors such that neighboring elements receive disjoint sets of colors. By using the definition of balanced and complete fractional stable we give for cubic circulant graphs an upper bounds on the fractional total chromatic number and for 4-regular circulant graphs we find the total chromatic number for some cases and we give the exact value of the fractional total chromatic number in most cases

L’évolution du web, de ce qui était typiquement connu comme un moyen de communication à sens unique en mode conversationnel, a radicalement changé notre manière de traiter l’information. Des sites de médias sociaux tels que Flickr et Facebook, offrent des espaces d’échange et de diffusion de l’information. Une information de plus en plus riche, mais aussi personnelle, et qui s’organise, le plus souvent, autour d’événements de la vie réelle. Ainsi, un événement peut être perçu comme un ensemble de vues personnelles et locales, capturées par différents utilisateurs. Identifier ces différentes instances permettrait, dès lors, de reconstituer une vue globale de l’événement. Plus particulièrement, lier différentes instances d’un même événement profiterait à bon nombre d’applications tel que la recherche, la navigation ou encore le filtrage et la suggestion de contenus. L’objectif principal de cette thèse est l’identification du contenu multimédia, associé à un événement dans de grandes collections d’images. Une première contribution est une méthode de recherche d’événements basée sur le contenu visuel. La deuxième contribution est une approche scalable et distribuée pour la construction de graphes des K plus proches voisins. La troisième contribution est une méthode collaborative pour la sélection de contenu pertinent. Plus particulièrement, nous nous intéresserons aux problèmes de génération automatique de résumés d’événements et suggestion de contenus dans les médias sociaux
The ease of publishing content on social media sites brings to the Web an ever increasing amount of user generated content captured during, and associated with, real life events. Social media documents shared by users often reflect their personal experience of the event. Hence, an event can be seen as a set of personal and local views, recorded by different users. These event records are likely to exhibit similar facets of the event but also specific aspects. By linking different records of the same event occurrence we can enable rich search and browsing of social media events content. Specifically, linking all the occurrences of the same event would provide a general overview of the event. In this dissertation we present a content-based approach for leveraging the wealth of social media documents available on the Web for event identification and characterization. To match event occurrences in social media, we develop a new visual-based method for retrieving events in huge photocollections, typically in the context of User Generated Content. The main contributions of the thesis are the following : (1) a new visual-based method for retrieving events in photo collections, (2) a scalable and distributed framework for Nearest Neighbors Graph construction for high dimensional data, (3) a collaborative content-based filtering technique for selecting relevant social media documents for a given event

Nous définissons et étudions un problème d'optimisation combinatoire et un programme linéaire en nombres entiers, qui modélise le routage multi-chemin dans un réseau sur puce à garantie de trafic. Basé sur le multiplexage temporel et l'émission cyclique des messages, le modèle permet d'éviter les collisions, les blocages statiques et dynamiques dans des réseaux à topologie irrégulière, tout en minimisant les temps de latence. Une extension de ce problème de routage multi-chemin, qui permet une reconfiguration dynamique du routage au moment de l'exécution est également présentée. Dans ce cas, des ensembles indépendants de chemins valides sont pré-calculés de telle sorte qu'ils peuvent être inter-changés en cours d'exécution sans impact sur le trafic courant, tout en réutilisant tous les intervalles de temps dont les ressources sont vacantes ou libérées. L'approche du graphe spatio-temporel étendu est retenue dans les processus de résolution. Tout d'abord, nous présentons un ensemble d'opérateurs de base de calcul de plus courts chemins. Se sont une heuristique de construction parallèle gloutonne, un opérateur de voisinage, et un algorithme de Dijkstra modifié dans un graphe spatio-temporel étendu qui calcul un chemin unique dans un NoC occupé en temps pseudo-polynomial. Ensuite, pour résoudre l'ensemble des problèmes, les opérateurs sont introduits et combinés dans trois méthodes de recherche locale itérée capable de générer rapidement des solutions admissibles, un algorithme évolutionnaire à base de population solutions conférant une grande diversité à la recherche de solutions et un algorithme mémétique, tirant partie des avantages des deux précédents. Les expériences sont réalisées sur un ensemble d'instances d'applications réelles, et d'instances d'applications artificielles générées aléatoirement à partir des cas réels, pour illustrer les performances et la robustesse des méthodes de recherche
We define and study a combinatorial optimization problem and mixed-integer linear programming, that models multi-path routing in a Network-on-Chip with guaranteed traffic. Based on time division multiplexing, the model allows to avoid collisions, deadlocks and livelocks in irregular network topologies, while minimizing latency. An extension of this multi-path problem is also presented that allows dynamic reconfigurable routing at execution time. In that case, independent sets of valid routes are pre-computed in such a way they can be interchanged during execution with no impact on the existing traffic, while reusing all the vacant and free time-slot resources. A time-expanded graph approach is retained for the solution process. First, we present a set of basic operators to compute shortest paths. They can be a greedy parallel construction heuristic, neighborhood operators, and a modified Dijkstra algorithm in a time expanded graph that allows computing a single path in an occupied Noc in pseudo-polynomial time. Then, to solve all the problems, operators are introduced and combined within three iterated local search methods that can quickly generate feasible solutions, an evolutionary algorithm based on population conferring diversity solutions in search of solutions and a memetic algorithm, taking advantage of the benefits of the previous two. Experiments are done on a set of benchmarks representative of real life applications, and instances of artificial applications randomly generated from real cases, to illustrate the performance and robustness of research methods

L’évolution du web, de ce qui était typiquement connu comme un moyen de communication à sens unique en mode conversationnel, a radicalement changé notre manière de traiter l’information. Des sites de médias sociaux tels que Flickr et Facebook, offrent des espaces d’échange et de diffusion de l’information. Une information de plus en plus riche, mais aussi personnelle, et qui s’organise, le plus souvent, autour d’événements de la vie réelle. Ainsi, un événement peut être perçu comme un ensemble de vues personnelles et locales, capturées par différents utilisateurs. Identifier ces différentes instances permettrait, dès lors, de reconstituer une vue globale de l’événement. Plus particulièrement, lier différentes instances d’un même événement profiterait à bon nombre d’applications tel que la recherche, la navigation ou encore le filtrage et la suggestion de contenus. L’objectif principal de cette thèse est l’identification du contenu multimédia, associé à un événement dans de grandes collections d’images. Une première contribution est une méthode de recherche d’événements basée sur le contenu visuel. La deuxième contribution est une approche scalable et distribuée pour la construction de graphes des K plus proches voisins. La troisième contribution est une méthode collaborative pour la sélection de contenu pertinent. Plus particulièrement, nous nous intéresserons aux problèmes de génération automatique de résumés d’événements et suggestion de contenus dans les médias sociaux
The ease of publishing content on social media sites brings to the Web an ever increasing amount of user generated content captured during, and associated with, real life events. Social media documents shared by users often reflect their personal experience of the event. Hence, an event can be seen as a set of personal and local views, recorded by different users. These event records are likely to exhibit similar facets of the event but also specific aspects. By linking different records of the same event occurrence we can enable rich search and browsing of social media events content. Specifically, linking all the occurrences of the same event would provide a general overview of the event. In this dissertation we present a content-based approach for leveraging the wealth of social media documents available on the Web for event identification and characterization. To match event occurrences in social media, we develop a new visual-based method for retrieving events in huge photocollections, typically in the context of User Generated Content. The main contributions of the thesis are the following : (1) a new visual-based method for retrieving events in photo collections, (2) a scalable and distributed framework for Nearest Neighbors Graph construction for high dimensional data, (3) a collaborative content-based filtering technique for selecting relevant social media documents for a given event

Dans la plupart des systèmes nécessitant la modélisation de l'utilisateur pour adapter l'information à ses besoins spécifiques, l'utilisateur est représenté avec un profil généralement composé de ses centres d'intérêts. Les centres d'intérêts de l'utilisateur sont construits et enrichis au fil du temps à partir de ses interactions avec le système. De par cette nature évolutive des centres d'intérêts de l'utilisateur, le profil de l'utilisateur ne peut en aucun moment être considéré comme entièrement connu par un système. Cette connaissance partielle du profil de l'utilisateur à tout instant t a pour effet de réduire considérablement les performances des mécanismes d'adaptation de l'information à l'utilisateur lorsque le profil de l'utilisateur ne contient pas (ou contient très peu) les informations nécessaires à leur fonctionnement. Cet inconvénient est particulièrement plus récurrent chez les nouveaux utilisateurs d'un système (instant t=0, problème du démarrage à froid) et chez les utilisateurs peu actifs. Pour répondre à cette problématique, plusieurs travaux ont exploré des sources de données autres que celles produites par l'utilisateur dans le système : utilisateurs au comportement similaire (utilisé dans le filtrage collaboratif) ou données produites par l'utilisateur dans d'autres systèmes (conception de profil utilisateur multi-application et gestion des identités multiples des utilisateurs). Très récemment, avec l'avènement du Web social et l'explosion des réseaux sociaux en ligne, ces derniers sont de plus en plus étudiés comme source externe de données pouvant servir à l'enrichissement du profil de l'utilisateur. Ceci a donné naissance à de nouveaux mécanismes de filtrage social de l'information : systèmes de recherche d'information sociale, systèmes de recommandation sociaux, etc. Les travaux actuels portant sur les mécanismes de filtrage social de l'information démontrent que ce nouveau champ de recherche est très prometteur. Une étude sur les travaux existants nous permet tout de même de noter particulièrement deux faiblesses : d'une part, chacune des approches proposées dans ces travaux reste très spécifique à son domaine d'application (et au mécanisme associé), et d'autre part, ces approches exploitent de manière unilatérale les profils des individus autour de l'utilisateur dans le réseau social. Pour pallier ces deux faiblesses, nos travaux de recherche proposent une démarche méthodique permettant de définir d'une part un modèle social générique de profil de l'utilisateur réutilisable dans plusieurs domaines d'application et par différents mécanismes de filtrage social de l'information, et à proposer d'autre part, une technique permettant de dériver de manière optimale des informations du profil de l'utilisateur à partir de son réseau social. Nous nous appuyons sur des travaux existants en sciences sociales pour proposer une approche d'usage des communautés (plutôt que des individus) autour de l'utilisateur. La portion significative de son réseau social est constituée des individus situés à une distance maximum k de l'utilisateur et des relations entre ces individus (réseau k-égocentrique). A partir de deux évaluations de l'approche proposée, l'une dans le réseau social numérique Facebook, et l'autre dans le réseau de co-auteurs DBLP, nous avons pu démontrer la pertinence de notre approche par rapport aux approches existantes ainsi que l'impact de mesures telles que la centralité de communautés (degré ou proximité par exemple) ou la densité des réseaux k-égocentriques sur la qualité des résultats obtenus. Notre approche ouvre de nombreuses perspectives aux travaux s'intéressant au filtrage social de l'information dans de multiples domaines d'application aussi bien sur le Web (personnalisation de moteurs de recherche, systèmes de recommandation dans le e-commerce, systèmes adaptatifs dans les environnements e-Learning, etc. ) que dans les intranets d'entreprise (systèmes d'analyses comportementales dans les réseaux d'abonnés de clients télécoms, détection de comportements anormaux/frauduleux dans les réseaux de clients bancaires, etc. )
In most systems that require user modeling to adapt information to each user's specific need, a user is usually represented by a user profile in the form of his interests. These interests are learnt and enriched over time from users interactions with the system. By the evolving nature of user's interests, the user's profile can never be considered fully known by a system. This partial knowledge of the user profile at any time t significantly reduces the performance of adaptive systems, when the user's profile contains no or only some information. This drawback is particularly most recurrent for new users in a system (time t = 0, also called cold start problem) and for less active users. To address this problem, several studies have explored data sources other than those produced by the user in the system: activities of users with similar behavior (e. G. Collaborative filtering techniques) or data generated by the user in other systems (e. G. , multi-application user's profiles, multiple identities management systems). By the recent advent of Social Web and the explosion of online social networks sites, social networks are more and more studied as an external data source that can be used to enrich users' profiles. This has led to the emergence of new social information filtering techniques (e. G. Social information retrieval, social recommender systems). Current studies on social information filtering show that this new research field is very promising. However, much remains to be done to complement and enhance these studies. We particularly address two drawbacks: (i) each existing social information filtering approach is specific in its field scope (and associated mechanisms), (ii) these approaches unilaterally use profiles of individuals around the user in the social network to improve traditional information filtering systems. To overcome these drawbacks in this thesis, we aim at defining a generic social model of users' profiles that can be reusable in many application domains and for several social information filtering mechanisms, and proposing optimal techniques for enriching user's profile from the user's social network. We rely on existing studies in social sciences to propose a communities (rather than individuals) based approach for using individuals around the user in a specific part of his social network, to derive his social profile (profile that contains user's interest derived from his social network). The significant part of the user's social network used in our studies is composed of individuals located at a maximum distance k (in the entire social network) from the user, and relationships between these individuals (k-egocentric network). Two evaluations of the proposed approach based on communities in k-egocentric networks have been conducted in the online social network Facebook and the co-authors network DBLP. They allow us to demonstrate the relevance of the proposal with respect to existing individual based approaches, and the impact of structural measures such as the centrality of communities (degree or proximity) or user's k-egocentric network density, on the quality of results. Our approach opens up many opportunities for future studies in social information filtering and many application domains as well as on the Web (e. G. Personalization of search engines, recommender systems in e-commerce, adaptive systems in e-Learning environment) or in Intranets business systems (e. G. Behavioral analysis in networks of subscribers telecom customers, detection of abnormal behavior network bank customers, etc. )

So far, I have examined children’s spellings at the level of whole words. The results show that children have more difficulty with some kinds of words than others. For example, children often misspell words that contain multiple-letter graphemes, words such as that and sang. Children often misspell irregular words, words such as said and come. One would guess that th is the trouble spot in that and ai is the trouble spot in said. However, because the analyses presented so far are confined to whole words, I cannot say for sure. To determine which parts of words are difficult to spell, I must move from the level of whole words to the level of individual phonemes and individual graphemes. The need to examine children’s spellings at the level of phonemes and graphemes stems from the nature of the English writing system itself. As discussed in Chapter 1, the English writing system is basically alphabetic. Although most phonemes may be spelled in more than one way, there are relations between phonemes and graphemes. For instance, /k/ may be spelled with k, as in key, c, as in care, or ck, as in back, among other possibilities. Adults cannot always choose the correct spelling from among these possibilities, but we know that /k/ could never be written with m or b. Our knowledge of phoneme-grapheme correspondences tells us that Carl or Karl are reasonable renditions of the spoken form /k’arl/ but that Marl is not. Traditionally, it was thought that children learn to spell on a visual basis, by memorizing the sequence of letters in each word. In this view, children treat printed words as wholes. They do not learn relations between the parts of printed words (graphemes) and the parts of spoken words (phonemes). The traditional view further implies that children memorize one word at a time. They do not learn relations between sounds and spellings that apply to many different words. Findings reported in Chapter 2 suggest that this traditional view of learning to spell is incorrect For example, children's difficulty on irregular words like said and come suggests that children learn about the correspondences between phonemes and graphemes.

Learning to spell involves learning about the relations between the phonemes of the spoken language and the graphemes of the printed language. In Chapter 4, I asked how children learn these relations for vowels. The results showed that a number of factors affect children’s learning, including their exposure to printed words, their knowledge of letter names, and their phonological systems. In this chapter, I turn to consonants. I ask whether these same factors affect children’s spelling of consonants. This chapter focuses on substitution errors and, to a lesser extent, correct spellings. Consonant omission errors will be considered in detail in Chapter 8. Sometimes, the first graders’ most common spellings of consonant phonemes were those spellings that are most frequent in the conventional English system. However, the children’s spellings did not always mirror those of conventional English. The children sometimes used a grapheme that never represents the phoneme in the standard system; that is, an illegal spelling. As in Chapter 4, I focus on illegal spellings that occurred at rates of 2.5% or more. I ask why the children selected that particular grapheme to represent the phoneme. In other cases, the students used a legal spelling significantly more often than expected given its frequency in the conventional system. Again, factors other than exposure to the relations between phonemes and graphemes in English words must be responsible for the error. I ask what these factors are. As in Chapter 4, I use binomial tests to compare the frequencies of correspondences in children’s spelling to the frequencies of the correspondences in the conventional spellings of the same words. In this section, the children’s spellings of various consonant phonemes are discussed. The reader may find it helpful to refer to the consonant chart of Figure 1.5 when reading this section. The stop consonants of English are /p/, /t/, /k/, /b/, /d/, and /g/. In discussing how the children spelled these consonants, I will first consider the children’s spellings without regard to the contexts in which the consonants occurred. Next, I will discuss some errors that occurred for stop consonants in particular contexts.

This chapter focuses on organic and nanoscale superconducting systems. The tetramethyl-tetraselenafulvalene (TMTSF) and ethylenedithiotetrathiafulvalene (ET) organic families, along with the fullerides, are described. Special attention is paid to graphene-like structures, which are examples of two-dimensional systems. As for nanoscale systems, small-scale nanoscale structures are introduced and the pairing in aromatic molecules, like coronene, is discussed. The presence of the energy shell in some nanoclusters makes the pairing of electrons with opposite projections of orbital momenta perfectly realistic. This phenomenon has been observed experimentally for some aluminium clusters with a Tc on the order of 120 K. The nano-based tunnelling networks can transfer a macroscopic dissipationless current. Interface superconductivity is discussed, with a special focus on the FeSe/SrTiO₃ system. The dream of room temperature superconductivity, envisioned shortly after the discovery of the phenomenon, has become perfectly realistic. This final chapter on materials describes various paths towards to this goal.

In this work we investigate the effect of molecular alignment on thermal conductivity (k) enhancement of polyethylene/graphene nanoplatelet (PE/GNP) composites. Enhancement of thermal conductivity of polymers can pave way for their application in heat exchangers leading to significant energy savings as processing of polymers is more energy efficient than metals. Such energy savings will drive down costs and will have the additional benefit of considerably reducing the environmental effects of energy production. Such high k polymers will also enable improved thermal management in electronic devices in servers, automobiles, high brightness LEDs and mobile applications. Stretching is known to induce alignment of molecular chains in a polymer system thereby increasing thermal conductivity. In this work we explore mechanical stretching of polyethylene-graphene nanocomposites to enhance their k.

Fundamental understanding of the water in graphene is crucial to optimally design and operate the sustainable energy, water desalination, and bio-medical systems. A numerous atomic-scale studies have been reported, primarily articulating the surface interactions (interatomic potentials) between the water and graphene. However, a systematic comparative study among the various interatomic potentials is rare, especially for the water transport confined in the graphene nanostructure. In this study, the effects of different interatomic potentials and gap sizes on water self-diffusivity are investigated using the molecular dynamics simulation at T = 300 K. The water is confined in the rigid graphene nanogap with the various gap sizes Lz = 0.7 to 4.17 nm, using SPC/E and TIP3P water models. The water self-diffusivity is calculated using the mean squared displacement approach. It is found that the water self-diffusivity in the confined region is lower than that of the bulk water, and it decreases as the gap size decreases and the surface energy increases. Also, the water self-diffusivity nearly linearly decreases with the increasing surface energy to reach the bulk water self-diffusivity at zero surface energy. The obtained results provide a roadmap to fundamentally understand the water transport properties in the graphene geometries and surface interactions.

The preparation and physico-chemical characterization of mesoporous and microporous carbons, obtained via a casting procedure, from a SBA-15 silica and a commercial Na-Y zeolite, is reported. XRD spectra showed that ordered carbon replicas occur in all cases. Micro-Raman spectra showed that rather homogeneous powders are obtained, exhibiting the presence of a graphitized carbon phase of small imperfect graphene sheets, typical of sp2 C, along with an amorphous one, notwithstanding the relatively low temperature adopted during the carbonization processes (1173 K). N2 adsorption isotherms at 77 K allowed the determination of BET surface areas and pore volumes: on account of the high porosity and the low specific weight, with respect to zeolites, for example, these carbon materials could be promising media for hydrogen storage. They could be used as such, or after convenient functionalization or metal doping.

The nano-scale interaction between polymer molecules and nanoparticle is a key factor in determining the macro-scale strength of the composite. In recent years numerous efforts have been directed towards modeling nanocomposites in order to better understand the reasons behind the enhancement of mechanical properties, even by the slight addition (a few weight percent) of nano-materials. In order to better understand the local influence of nano-particle on the mechanical properties of the composite, it is required to perform nano-scale analysis. In this context, modeling of fracture and damage in nano-graphene reinforced EPON 862 has been discussed in the current paper. Regarding fracture in polymers, the critical value of the J-integral (JI), where the subscript I denotes the fracture mode (I=1, 2, 3), at crack initiation could be used as a suitable metric for estimating the crack driving force as well as fracture toughness of the material as the crack begins to initiate. However, for the conventional macroscale definition of the J-integral to be valid at the nanoscale, in terms of the continuum stress and displacement fields and their spatial derivatives — requires the construction of local continuum fields from discrete atomistic data, and using these data in the conventional contour integral expression for atomistic J-integral. One such methodology is proposed by Hardy that allows for the local averaging necessary to obtain the definition of free energy, deformation gradient, and Piola-Kirchoff stress as fields (and divergence of fields) and not just as total system averages. Further, the atomistic J-integral takes into account the effect of reduction in J from continuum estimates due to the fact that the free energy available for crack propagation is less than the internal energy at sufficiently high temperatures when entropic contributions become significant. In this paper, the proposed methodology is used to compute J-integral using atomistic data obtained from LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator). As a case study, the feasibility of computing the dynamic atomistic J-integral over the MD domain is evaluated for a graphene nano-platelet with a central crack using OPLS (Optimized Potentials for Liquid Simulations) potential. For model verification, the values of atomistic J-integral are compared with results from linear elastic fracture mechanics (LEFM) for isothermal crack initiation at 0 K and 300 K. Computational results related to the path-independence of the atomistic J-Integral are also presented. Further, a novel approach that circumvents the complexities of direct computation of entropic contributions is also discussed. Preliminary results obtained from the bond-order based ReaxFF potential for 0.1 K and 300 K are presented, and show good agreement with the predictions.

Due to the high intrinsic thermal conductivity of carbon allotropes, there have been many attempts to incorporate such structures into existing thermal abatement technologies. In particular, carbon nanotubes (CNTs) and graphitic materials (i.e., graphite and graphene flakes or stacks) have garnered much interest due to the combination of both their thermal and mechanical properties. However, the introduction of these carbon-based nanostructures into thermal abatement technologies greatly increases the number of interfaces per unit length within the resulting composite systems. Consequently, thermal transport in these systems is governed as much by the interfaces between the constituent materials as it is by the materials themselves. This paper reports the behavior of phononic thermal transport across interfaces between isotropic thin films and graphite substrates. Elastic and inelastic diffusive transport models are formulated to aid in the prediction of conductance at a metal-graphite interface. The temperature dependence of the thermal conductance at Au-graphite interfaces is measured via transient thermoreflectance from 78 to 400 K. It is found that different substrate surface preparations prior to thin film deposition have a significant effect on the conductance of the interface between film and substrate.

A unique way for maximizing turbulent free convection from heated vertical plates to cold gases is studied in this paper. The central idea is to examine the attributes that binary gas mixtures having helium as the principal gas and xenon, nitrogen, oxygen, carbon dioxide, methane, tetrafluoromethane and sulfur hexafluoride as secondary gases may bring forward. From fluid physics, it is known that the thermo-physical properties affecting free convection with binary gas mixtures are viscosity ηmix, thermal conductivity λmix, density ρmix, and heat capacity at constant pressure. The quartet ηmix, λmix, ρmix, and Cp,mix is represented by triple-valued functions of the film temperature the pressure P, and the molar gas composition w. The viscosity is obtained from the Kinetic Theory of Gases conjoined with the Chapman-Enskog solution of the Boltzmann Transport Equation. The thermal conductivity is computed from the Kinetic Theory of Gases. The density is determined with a truncated virial equation of state. The heat capacity at constant pressure is calculated from Statistical Thermodynamics merged with the standard mixing rule. Using the similarity variable method, the descriptive Navier-Stokes and energy equations for turbulent Grashof numbers Grx > 109 are transformed into a system of two nonlinear ordinary differential equations, which is solved by the shooting method and the efficient fourth-order Runge-Kutta-Fehlberg algorithm. The numerical temperature fields T(x, y) for the five binary gas mixtures He-Xe, He-N2, He-O2, He-CO2, He-CH4, He-CF4 and He-SF6 are channeled through the allied mean convection coefficient hmix/B varying with the molar gas composition w in proper w-domain [0, 1]. For the seven binary gas mixtures utilized, the allied mean convection coefficient hmix/B versus the molar gas composition w is graphed in congruous diagrams. At a low film temperature Tf = 300 K, the global maximum allied mean convection coefficient hmix,max/B = 85 is furnished by the He-SF6 gas mixture at an optimal molar gas composition wopt = 0.93. The global maximum allied mean convection coefficient hmix,max/B = 57 is supplied by pure methane gas SF6 (w = 1) at a high film temperature Tf = 1000 K instead of the He-SF6 gas mixture.

Golfers must hone the sequence of the golf swing for consistent and efficient results, where a good golf shot is both precise and accurate. It takes countless hours of practice to develop the skill, and methods such as video feedback and coaching serve as aids to the process. However, few methods outside of visual observations exist to identify factors in the swing that cause a poor shot. Most golf analysis equipment is expensive and requires extensive setup time. Some systems only measure the physics of the ball and are limited to the practice range, whereas wearable sensors for golf are limited to specific motions such as the wrist. To address these challenges, a fabric-based, on-body sensor was developed and investigated to assess biomechanical movements during the swing. The wearable sensor, herein referred to as Motion Tape (MT), is a low-profile, disposable, self-adhesive, skin-strain sensor formed by spray-coating piezoresistive graphene nanocomposites directly onto kinesiology tape (K-tape). The objective of this study is to use MT to identify key movements in the swing sequence at four body locations: wrist, flexor carpi, anterior deltoid, and torso. First, MT sensors were fabricated for testing. Second, a human subject test protocol recording the golf swing of an experienced golfer was designed and conducted with participants wearing four MT sensors at the aforementioned locations. Last, the test data were processed, and the results showed that MT was able to identify unique movements during the swing. The MT data was analyzed using machine learning algorithms to identify movement abnormalities associated with poor swings. This allows for analysis of swing tempo for direct feedback to the golfer for improved performance.