Interdisciplinary Research (IDR) has received a lot of attention from academics, policy-makers, and decision-makers alike. RCUK invests £3 billion in research grants each year (RCUK 2017); half of the grants are provided to investigators who hail from different departments. There is mounting awareness of the challenges facing IDR, and a large body of literature trying to establish how IDR can be analysed (Davidson 2015, Yegros-Yegros, Rafols et al. 2015). Of these, the majority have been qualitative studies and it has been noticed that there is a distinct lack of quantitative studies that can be used to identify how to enable IDR. The literature shows that many of the barriers to IDR can be classified as either cultural or administrative (Katz and Martin 1997, Cummings and Kiesler 2005, Rafols 2007, Wagner, Roessner et al. 2011), neither of which are easily changed over a short period of time. The perspective taken in this research is that change can be affected by enabling the individuals who conduct IDR. Herein lies the main challenge; how can these future leaders of IDR be identified so that they can be properly supported. No existing datasets were deemed suitable for the purpose, and a new dataset was created to analyse IDR. To isolate dynamics within an organisation, hard boundaries were drawn around research-organisations. The University of Bath journal co-authorship dataset 2000-2017 was determined to be suitable for this purpose. From this dataset a co-authorship network was created. To analyse this, established models from literature were adapted and used to identify differences in disciplinary and interdisciplinary archetypes. This was done through a correlational study. No statistically significant differences between such author archetypes were found. It was therefore concluded that an alternative approach was necessary. By adapting the networks framework to account for different types of links between edges, a multilayer perspective was adopted. This resulted in a rank-3 tensor, node-aligned framework being proposed, allowing disciplines to be represented in the network. By using this framework to construct the University of Bath multiplex co-authorship network, an exemplar structure was established through use of a series of proposed structural metrics. A growth model was proposed and successfully recreated the structure and thereby uncovered mechanics affecting real-world multiplex networks. This highlighted the importance of node entities and the layer closeness centrality. This implies that it is very difficult to carry over benefits across disciplines, and that some disciplines are better suited to share and adapt knowledge than others. The growth model also allowed an analytical expression for the rate of change of disciplinary degree, thereby providing a model for who is most likely to enable and sustain IDR.

Doctor of Philosophy / Department of Electrical and Computer Engineering / Caterina Scoglio / Society increasingly depends on networks for almost every aspect of daily life. Over the past decade, network science has flourished tremendously in understanding, designing, and utilizing networks. Particularly, network science has shed light on the role of the underlying network topology on the dynamic behavior of complex systems, including cascading failure in power-grids, financial contagions in trade market, synchronization, spread of social opinion and trends, product adoption and market penetration, infectious disease pandemics, outbreaks of computer worms, and gene mutations in biological networks. In the last decade, most studies on complex networks have been confined to a single, often homogeneous network. An extremely challenging aspect of studying these complex systems is that the underlying networks are often heterogeneous, composite, and interdependent with other networks. This challenging aspect has very recently introduced a new class of networks in network science, which we refer to as multilayer and interconnected networks. Multilayer networks are an abstract representation of interconnection among nodes representing individuals or agents, where the interconnection has a multiple nature. For example, while a disease can propagate among individuals through a physical contact network, information can propagate among the same individuals through an online information-dissemination network. Another example is viral information dissemination among users of online social networks; one might disseminate information received from a Facebook contact to his or her followers on Twitter. Interconnected networks are abstract representations where two or more simple networks, possibly with different dynamics over them, are interconnected to each other. For example, in zoonotic diseases, a virus can move from the network of animals, with some transmission dynamics, to a human network, with possibly very different dynamics. As communication systems are evolving more and more toward integration with computing, sensing, and control systems, the theory of multilayer and interconnected networks seems to be crucial to successful communication systems development in cyber-physical infrastructures. Among the most relevant dynamics over networks is epidemic spreading. Epidemic spreading dynamics over simple networks exhibit a clear example where interaction between non-complex dynamics at node level and the topology leads to a complex emergent behavior. A substantial line of research during the past decade has been devoted to capturing the role of the network on spreading dynamics, and mathematical tools such as spectral graph theory have been greatly useful for this goal. For example, when the network is a simple graph, the dominant eigenvalue and eigenvector of the adjacency matrix have been proven to be key elements determining spreading dynamics features, including epidemic threshold, centrality of nodes, localization of spreading sites, and behavior of the epidemic model close to the threshold. More generally, for many other dynamics over a single network, dependency of dynamics on spectral properties of the adjacency matrix, Laplacian matrix, or some other graph-related matrix, is well-studied and rigorously established, and practical applications have been successfully derived. In contrast, limited established results exist for dynamics on multilayer and interconnected networks. Yet, an understanding of spreading processes over these networks is very important to several realistic phenomena in modern integrated and composite systems, including cascading failure in power grids, financial contagions in trade market, synchronization, spread of social opinion and trends, product adoption and market penetration, infectious disease pandemics, and outbreak in computer worms. This dissertation focuses on spreading processes on multilayer and interconnected networks, organized in three parts. The first part develops a general framework for modeling epidemic spreading in interconnected and multilayer networks. The second part solves two fundamental problems: introducing the concept of an epidemic threshold curve in interconnected networks, and coexistence phenomena in competitive spreading over multilayer networks. The third part of this dissertation develops an epidemic model incorporating human behavior, where multi-layer network formulation enables modeling and analysis of important features of human social networks, such as an information-dissemination network, as well as contact adaptation. Finally, I conclude with some open research directions in the topic of spreading processes over multilayer and interconnected networks, based on the resulting developments of this dissertation.

Master of Science / Department of Electrical and Computer Engineering / Caterina M. Scoglio / The generalized epidemic modeling framework simulator (GEMFsim) is a tool designed by Dr. Faryad Sahneh, former PhD student in the NetSE group. GEMFsim simulates stochastic spreading process over complex networks. It was first introduced in Dr. Sahneh’s doctoral dissertation "Spreading processes over multilayer and interconnected networks" and implemented in Matlab. As limited by Matlab language, this implementation typically solves only small networks; the slow simulation speed is unable to generate enough results in reasonable time for large networks. As a generalized tool, this framework must be equipped to handle large networks and contain sufficient support to provide adequate performance. The C language, a low-level language that effectively maps a program to machine in- structions with efficient execution, was selected for this study. Following implementation of GEMFsim in C, I packed it into Python and R libraries, allowing users to enjoy the flexibility of these interpreted languages without sacrificing performance. GEMFsim limitations are not limited to language, however. In the original algorithm (Gillespie’s Direct Method), the performance (simulation speed) is inversely proportional to network size, resulting in unacceptable speed for very large networks. Therefore, this study applied the Next Reaction Method, making the performance irrelevant of network size. As long as the network fits into memory, the speed is proportional to the average node degree of the network, which is not very large for most real-world networks. This study also applied parallel computing in order to advantageously utilize multiple cores for repeated simulations. Although single simulation can not be paralleled as a Markov process, multiple simulations with identical network structures were run simultaneously, sharing one network description in memory.

Online social networks have become a powerful venue for political activism. In many cases large, insular online communities form that have been shown to be powerful diffusion mechanisms of both misinformation and propaganda. In some cases these groups users advocate actions or policies that could be construed as extreme along nearly any distribution of opinion, and are thus called Online Extremist Communities (OECs). Although these communities appear increasingly common, little is known about how these groups form or the methods used to influence them. The work in this thesis provides researchers a methodological framework to study these groups by answering three critical research questions: How can we detect large dynamic online activist or extremist communities? What automated tools are used to build, isolate, and influence these communities? What methods can be used to gain novel insight into large online activist or extremist communities? These group members social ties can be inferred based on the various affordances offered by OSNs for group curation. By developing heterogeneous, annotated graph representations of user behavior I can efficiently extract online activist discussion cores using an ensemble of unsupervised machine learning methods. I call this technique Ensemble Agreement Clustering. Through manual inspection, these discussion cores can then often be used as training data to detect the larger community. I present a novel supervised learning algorithm called Multiplex Vertex Classification for network bipartition on heterogeneous, annotated graphs. This methodological pipeline has also proven useful for social botnet detection, and a study of large, complex social botnets used for propaganda dissemination is provided as well. Throughout this thesis I provide Twitter case studies including communities focused on the Islamic State of Iraq and al-Sham (ISIS), the ongoing Syrian Revolution, the Euromaidan Movement in Ukraine, as well as the alt-Right.

Mathematical modeling of spreading processes have been largely studied in the literature, and its presented a boom in the past few years. This is a fundamental task on the understanding and prediction of real spreading processes on top of a population and are subject to many structural and dynamical constraints. Aiming at a better understanding of this processes, we focused in two task: the modeling and the analysis of both dynamical and structural aspects of these processes. Initially, we proposed a new and general model that unifies epidemic and rumor spreading. Besides, regarding the analysis of these processes, we extended the classical formalism to multilayer networks, in which the theory was lacking. Interestingly, this study opened up new challenges concerning the understanding of multilayer networks. More specifically, regarding their spectral properties. In this thesis, we analyzed such processes on top of single and multilayer networks. Thus, throughout our analysis, we followed three complementary approaches: (i) analytical, (ii) numerical and (iii) simulations, mainly Monte Carlo simulations. Our main results are: (i) a new unifying model, enabling us to model and understand spreading processes on large systems, (ii) characterization of new phenomena on multilayer networks, such as layer-wise localization and the barrier effect and (iii) an spectral analysis of multilayer systems, suggesting a universal parameter and proposing a new analytical tool for its analysis. Our contributions enable further research on modeling of spreading processes, also emphasizing the importance of considering the complete multilayer structure instead of any coarse-graining. Additionally, it can be directly applied on the prediction and modeling real processes. Thus, aside from the theoretical interest and its mathematical implications, it also presents important social impact. / A modelagem matemática dos processos de disseminação tem sido amplamente estudada na literatura, sendo que o seu estudo apresentou um boom nos últimos anos. Esta é uma tarefa fundamental na compreensão e previsão de epidemias reais e propagação de rumores numa população, ademais, estas estão sujeitas a muitas restrições estruturais e dinâmicas. Com o objetivo de entender melhor esses processos, nos concentramos em duas tarefas: a de modelagem e a de análise de aspectos dinâmicos e estruturais. No primeiro, propomos um modelo novo e geral que une a epidemia e propagação de rumores. Também, no que diz respeito à análise desses processos, estendemos o formalismo clássico às redes multicamadas, onde tal teoria era inexistente. Curiosamente, este estudo abriu novos desafios relacionados à compreensão de redes multicamadas, mais especificamente em relação às suas propriedades espectrais. Nessa tese, analisamos esses processos em redes de uma e múltiplas camadas. Ao longo de nossas análises seguimos três abordagens complementares: (i) análises analíticas, (ii) experimentos numéricos e (iii) simulações de Monte Carlo. Assim, nossos principais resultados são: (i) um novo modelo que unifica as dinâmicas de rumor e epidemias, nos permitindo modelar e entender tais processos em grandes sistemas, (ii) caracterização de novos fenômenos em redes multicamadas, como a localização em camadas e o efeito barreira e (iii) uma análise espectral de sistemas multicamadas, sugerindo um parâmetro de escala universal e propondo uma nova ferramenta analítica para sua análise. Nossas contribuições permitem que novas pesquisas sobre modelagem de processos de propagação, enfatizando também a importância de se considerar a estrutura multicamada. Dessa forma, as nossas contribuições podem ser diretamente aplicadas à predição e modelagem de processos reais. Além do interesse teórico e matemático, nosso trabalho também apresenta implicações sociais importantes.

Aus den Mängeln bisher verwendeter Modelle zur Abbildung des Wellenablaufes zu Prognosezwecken im Hochwasserfall wird in dieser Arbeit eine Methodik entwickelt, die die Schnelligkeit und Robustheit künstlicher neuronaler Netze mit der Zuverlässigkeit hydrodynamisch-numerischer Modellierung verbindet. Ein eindimensionales hydrodynamisches Modell beinhaltet die genaue Kenntnis der Geometrie des Flußlaufes und der Vorländer und berücksichtigt die physikalischen Prozesse des Wellenablaufes. Mit diesem deterministischen Modell ist eine Grundlage für umfangreiche Szenarienrechnungen zur Erstellung einer Datenbasis geschaffen, die die weite Spanne theoretisch möglicher Hochwasserereignisse abdeckt. Mit dieser Datenbasis können dann künstliche neuronale Netze trainiert werden, die auch im Bereich extremer Hochwasserereignisse zuverlässige Prognosen liefern. In dieser Arbeit werden mit Multilayer-Feedforward-Netzen und selbstorganisierenden Merkmalskarten zwei Netztypen als Vertreter überwacht und unüberwacht lernender neuronaler Netze auf ihre diesbezügliche Eignung untersucht und beurteilt. Desweiteren wurde die Methodik auf die Einbeziehung von Merkmalen für die Niederschlags-Abfluß-Prozesse im unbeobachteten Zwischengebiet zur Berücksichtigung lateraler Zuflüsse entlang der modellierten Fließstrecken erweitert. Die Datenbasis wurde hierfür mit einem Niederschlags-Abfluß-Modell erstellt. Ein Hauptschwerpunkt liegt in der Überführung der Eingangsdaten in charakteristische Merkmale zur Abbildung der Zielgrößen, in diesem Falle des Durchflusses und Wasserstandes am Zielpegel. So dienen die deterministischen Modelle nicht nur zur Erstellung einer verläßlichen Datenbasis für das Training der Netze, sondern ermöglichen – sowohl für die Niederschlags-Abfluß-Prozesse, als auch für die hydrodynamischen Prozesse – Analysen betreffs der Sensitivität der Modellergebnisse infolge von Änderungen der Inputdaten. Mit Hilfe dieser Analysen werden wichtige Informationen zur Findung der relevanten Merkmale erlangt. Ein Schlüssel für die erfolgreiche Eingliederung der Niederschlags-Abfluß-Prozesse in das Prognosenetz ist die Einführung eines einzigen Zustandsmerkmals, welches die gesamte meteorologische Vorgeschichte des Ereignisses zur Charakterisierung des Gebietszustandes vereinigt. Die entwickelte Methodik wurde anhand des Einzugsgebietes der Freiberger Mulde erfolgreich getestet.

Connectomics – the study of neural connectivity – is primarily concerned with the mapping and characterisation of wired synaptic links; however, it is well established that long-distance chemical signalling via extrasynaptic volume transmission is also critical to brain function. As these interactions are not visible in the physical structure of the nervous system, current approaches to connectomics are unable to capture them. This work addresses the problem of missing extrasynaptic interactions by demonstrating for the first time that whole-animal volume transmission networks can be mapped from gene expression and ligand-receptor interaction data, and analysed as part of the connectome. Complete networks are presented for the monoamine systems of Caenorhabditis elegans, along with a representative sample of selected neuropeptide systems. A network analysis of the synaptic (wired) and extrasynaptic (wireless) connectomes is presented which reveals complex topological properties, including extrasynaptic rich-club organisation with interconnected hubs distinct from those in the synaptic and gap junction networks, and highly significant multilink motifs pinpointing locations in the network where aminergic and neuropeptide signalling is likely to modulate synaptic activity. Thus, the neuronal connectome can be modelled as a multiplex network with synaptic, gap junction, and neuromodulatory layers representing inter-neuronal interactions with different dynamics and polarity. This represents a prototype for understanding how extrasynaptic signalling can be integrated into connectomics research, and provides a novel dataset for the development of multilayer network algorithms.

Submitted by Clebson Anjos (clebson.leandro54@gmail.com) on 2016-02-16T22:43:34Z No. of bitstreams: 1 arquivototal.pdf: 6316800 bytes, checksum: 0b88bad394fd98379d16f9b515a3f984 (MD5) / Made available in DSpace on 2016-02-16T22:43:34Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 6316800 bytes, checksum: 0b88bad394fd98379d16f9b515a3f984 (MD5) Previous issue date: 2015-08-31 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The communication between robots is essential and of great interest in the robotics research eld. It is from it that the robots can discover and pass on information about obstacles in their way, pass on to the other robots information about a target that other robot can not detect, determine together which is the best route to follow, among others, until the goals are achieved. The major di culty in the context where robot communication exists in an Ad Hoc network is distancing the robots from each other. So that one of them for example, may migrate to another area leaving their initial formation, completing a speci c task and return to the starting point, in addition to distancing the whole formation of robots from a central computer. By distancing a robot from the limits imposed by other means will entail a loss of signal compromising communication between them. This way, knowing a wireless network coverage area where robots will be connected is an important process for the implementation of a system of communication between robots. The aim of this work is to demonstrate the feasibility of expanding the coverage area of a wireless network, to be used in multi-robot systems, of a formation of robots that use a Real Time Database (RTDB) in multilayer networks that using the IEEE 802.11 and 802.15.4, analyzing the behavior of this expansion in relation to connectivity of robots, coverage area and interference in wireless networks. According to the performed experiments, the results demonstrate the robustness and reliability of the proposed scenario and it was shown that the use of multilayer networks provides a greater range in the coverage area of a robot training. / A comunicação entre robôs é tarefa fundamental e de grande importância na robótica. É a partir dela que robôs conseguem descobrir e repassar informações sobre obstáculos encontrados no caminho, transmitir informações sobre um alvo que outro robô não consiga detectar ou determinar, em conjunto com outros robôs, a melhor rota a ser seguida, entre outros. Uma das grandes dificuldades, no contexto em que se tem a comunicação de robôs em uma rede sem o, está em distanciá-los. Um deles, por exemplo, pode migrar para uma outra área a m de realizar uma tarefa específi ca, distanciando-se, assim, da sua formação inicial e ultrapassando o limiar de cobertura de rede. Ao distanciar um robô do outro, os limites conferidos pelo meio acarretarão uma perda de sinal comprometendo a comunicação entre eles. Deste modo, conhecer a área de cobertura de uma rede sem o onde robôs estarão conectados é um processo importante para a implantação de um sistema de comunicação entre robôs. Este trabalho propõe expansão da área de cobertura de uma rede sem o, a ser utilizada em sistemas multi-robôs, usando um Middleware de Base de Dados em Tempo Real - RTDB, em redes multicamadas que utilizam os padrões IEEE 802.11 e 802.15.4, analisando o comportamento dessa expansão em relação a conectividade dos robôs, área de cobertura e de interferência de Redes Sem Fio. Os resultados obtidos demonstram a robustez e con fiabilidade do cenário proposto e de acordo com os experimentos realizados, comprovou-se que o uso de redes multicamadas proporciona um maior alcance na área de cobertura de uma formação de robôs.

Mathematical modeling of spreading processes have been largely studied in the literature, and its presented a boom in the past few years. This is a fundamental task on the understanding and prediction of real spreading processes on top of a population and are subject to many structural and dynamical constraints. Aiming at a better understanding of this processes, we focused in two task: the modeling and the analysis of both dynamical and structural aspects of these processes. Initially, we proposed a new and general model that unifies epidemic and rumor spreading. Besides, regarding the analysis of these processes, we extended the classical formalism to multilayer networks, in which the theory was lacking. Interestingly, this study opened up new challenges concerning the understanding of multilayer networks. More specifically, regarding their spectral properties. In this thesis, we analyzed such processes on top of single and multilayer networks. Thus, throughout our analysis, we followed three complementary approaches: (i) analytical, (ii) numerical and (iii) simulations, mainly Monte Carlo simulations. Our main results are: (i) a new unifying model, enabling us to model and understand spreading processes on large systems, (ii) characterization of new phenomena on multilayer networks, such as layer-wise localization and the barrier effect and (iii) an spectral analysis of multilayer systems, suggesting a universal parameter and proposing a new analytical tool for its analysis. Our contributions enable further research on modeling of spreading processes, also emphasizing the importance of considering the complete multilayer structure instead of any coarse-graining. Additionally, it can be directly applied on the prediction and modeling real processes. Thus, aside from the theoretical interest and its mathematical implications, it also presents important social impact. / A modelagem matemática dos processos de disseminação tem sido amplamente estudada na literatura, sendo que o seu estudo apresentou um boom nos últimos anos. Esta é uma tarefa fundamental na compreensão e previsão de epidemias reais e propagação de rumores numa população, ademais, estas estão sujeitas a muitas restrições estruturais e dinâmicas. Com o objetivo de entender melhor esses processos, nos concentramos em duas tarefas: a de modelagem e a de análise de aspectos dinâmicos e estruturais. No primeiro, propomos um modelo novo e geral que une a epidemia e propagação de rumores. Também, no que diz respeito à análise desses processos, estendemos o formalismo clássico às redes multicamadas, onde tal teoria era inexistente. Curiosamente, este estudo abriu novos desafios relacionados à compreensão de redes multicamadas, mais especificamente em relação às suas propriedades espectrais. Nessa tese, analisamos esses processos em redes de uma e múltiplas camadas. Ao longo de nossas análises seguimos três abordagens complementares: (i) análises analíticas, (ii) experimentos numéricos e (iii) simulações de Monte Carlo. Assim, nossos principais resultados são: (i) um novo modelo que unifica as dinâmicas de rumor e epidemias, nos permitindo modelar e entender tais processos em grandes sistemas, (ii) caracterização de novos fenômenos em redes multicamadas, como a localização em camadas e o efeito barreira e (iii) uma análise espectral de sistemas multicamadas, sugerindo um parâmetro de escala universal e propondo uma nova ferramenta analítica para sua análise. Nossas contribuições permitem que novas pesquisas sobre modelagem de processos de propagação, enfatizando também a importância de se considerar a estrutura multicamada. Dessa forma, as nossas contribuições podem ser diretamente aplicadas à predição e modelagem de processos reais. Além do interesse teórico e matemático, nosso trabalho também apresenta implicações sociais importantes.

Aus den Mängeln bisher verwendeter Modelle zur Abbildung des Wellenablaufes zu Prognosezwecken im Hochwasserfall wird in dieser Arbeit eine Methodik entwickelt, die die Schnelligkeit und Robustheit künstlicher neuronaler Netze mit der Zuverlässigkeit hydrodynamisch-numerischer Modellierung verbindet. Ein eindimensionales hydrodynamisches Modell beinhaltet die genaue Kenntnis der Geometrie des Flußlaufes und der Vorländer und berücksichtigt die physikalischen Prozesse des Wellenablaufes. Mit diesem deterministischen Modell ist eine Grundlage für umfangreiche Szenarienrechnungen zur Erstellung einer Datenbasis geschaffen, die die weite Spanne theoretisch möglicher Hochwasserereignisse abdeckt. Mit dieser Datenbasis können dann künstliche neuronale Netze trainiert werden, die auch im Bereich extremer Hochwasserereignisse zuverlässige Prognosen liefern. In dieser Arbeit werden mit Multilayer-Feedforward-Netzen und selbstorganisierenden Merkmalskarten zwei Netztypen als Vertreter überwacht und unüberwacht lernender neuronaler Netze auf ihre diesbezügliche Eignung untersucht und beurteilt. Desweiteren wurde die Methodik auf die Einbeziehung von Merkmalen für die Niederschlags-Abfluß-Prozesse im unbeobachteten Zwischengebiet zur Berücksichtigung lateraler Zuflüsse entlang der modellierten Fließstrecken erweitert. Die Datenbasis wurde hierfür mit einem Niederschlags-Abfluß-Modell erstellt. Ein Hauptschwerpunkt liegt in der Überführung der Eingangsdaten in charakteristische Merkmale zur Abbildung der Zielgrößen, in diesem Falle des Durchflusses und Wasserstandes am Zielpegel. So dienen die deterministischen Modelle nicht nur zur Erstellung einer verläßlichen Datenbasis für das Training der Netze, sondern ermöglichen – sowohl für die Niederschlags-Abfluß-Prozesse, als auch für die hydrodynamischen Prozesse – Analysen betreffs der Sensitivität der Modellergebnisse infolge von Änderungen der Inputdaten. Mit Hilfe dieser Analysen werden wichtige Informationen zur Findung der relevanten Merkmale erlangt. Ein Schlüssel für die erfolgreiche Eingliederung der Niederschlags-Abfluß-Prozesse in das Prognosenetz ist die Einführung eines einzigen Zustandsmerkmals, welches die gesamte meteorologische Vorgeschichte des Ereignisses zur Charakterisierung des Gebietszustandes vereinigt. Die entwickelte Methodik wurde anhand des Einzugsgebietes der Freiberger Mulde erfolgreich getestet.