In the context of a DFG research project, scientists of Prof. Helbing’s chair at the Institute of Transport & Economics deal with the dynamics of disasters, being experienced in the modelling of complex systems and in the simulation of emergency scenarios. The analyses of systems and their behaviour in extraordinary events are based on the latest results of network sciences and on numerous empirical investigations. The results shall be used for precaution measures and innovations in disaster recovery. / Im Rahmen eines Projekts der Deutschen Forschungsgemeinschaft (DFG-Projekt He 2789/6-1) befassen sich Wissenschaftler unter Leitung von Professor Dirk Helbing an der Fakultät Verkehrswissenschaften „Friedrich List“ mit der Dynamik von Katastrophen. Aus der Simulation von Fußgängerströmen, des Panikverhaltens von Menschen sowie der Verkehrsmodellierung verfügen die Wissenschaftler bereits über einschlägige Erfahrungen auf dem Gebiet der Modellierung komplexer Systeme sowie auf dem Gebiet der Simulation und Auswertung von Notfallszenarien. Auf Basis der jüngsten Ergebnisse der Netzwerkforschung und umfangreicher empirischer Untersuchungen von Katastrophenereignissen werden im Rahmen des aktuellen Forschungsprojekts anthropogene Systeme auf ihr Verhalten bei außergewöhnlichen Schadensereignissen untersucht. Die Projektergebnisse (Laufzeit bis 2007) sollen als Basis für weitere Verbesserungen in der Vorsorge und im Management von Katastrophen dienen.

Models of Public Mass Transportation in GIS Environment Abstract The thesis tries to describe and categorize models of public mass transportation. A new model is built upon these models and based on authors' thoughts. This model is constructed to work with headways only instead of timetables. Therefore one cannot precisely estimate the passengers' behaviour. The model works also with street network aside from public transportation lines. An application is built upon these model and allows network editing and processing of shortest-path analysis and accessibility analysis. Technologies used: ExtJS, OpenLayers, CGI, Shapely, STORM, PostgreSQL. This application is then used for building a public transportation network of Prague 6 region and for performing analysis upon this network.

In recent years, deep learning has made a significant impact in various fields – helping to push the state-of-the-art forward in many application domains. Convolutional Neural Networks (CNN) have been applied successfully to tasks such as visual object detection, image super-resolution, and video action recognition while Long Short-term Memory (LSTM) and Transformer networks have been used to solve a variety of challenging tasks in natural language processing. However, these popular deep learning architectures (i.e., CNNs, LSTMs, and Transformers) can only handle data that can be represented as grids or sequences. Due to this limitation, many existing deep learning approaches do not generalize to problem domains where the data is represented as graphs – social networks in social network analysis or molecular graphs in chemoinformatics, for instance. The goal of this thesis is to help bridge the gap by studying deep learning solutions that can handle graph data naturally. In particular, we explore deep learning-based approaches in the following areas. 1. Graph Attention. In the real-world, graphs can be both large – with many complex patterns – and noisy which can pose a problem for effective graph mining. An effective way to deal with this issue is to use an attention-based deep learning model. An attention mechanism allows the model to focus on task-relevant parts of the graph which helps the model make better decisions. We introduce a model for graph classification which uses an attention-guided walk to bias exploration towards more task-relevant parts of the graph. For the task of node classification, we study a different model – one with an attention mechanism which allows each node to select the most task-relevant neighborhood to integrate information from. 2. Graph Representation Learning. Graph representation learning seeks to learn a mapping that embeds nodes, and even entire graphs, as points in a low-dimensional continuous space. The function is optimized such that the geometric distance between objects in the embedding space reflect some sort of similarity based on the structure of the original graph(s). We study the problem of learning time-respecting embeddings for nodes in a dynamic network. 3. Brain Network Discovery. One of the fundamental tasks in functional brain analysis is the task of brain network discovery. The brain is a complex structure which is made up of various brain regions, many of which interact with each other. The objective of brain network discovery is two-fold. First, we wish to partition voxels – from a functional Magnetic Resonance Imaging scan – into functionally and spatially cohesive regions (i.e., nodes). Second, we want to identify the relationships (i.e., edges) between the discovered regions. We introduce a deep learning model which learns to construct a group-cohesive partition of voxels from the scans of multiple individuals in the same group. We then introduce a second model which can recover a hierarchical set of brain regions, allowing us to examine the functional organization of the brain at different levels of granularity. Finally, we propose a model for the problem of unified and group-contrasting edge discovery which aims to discover discriminative brain networks that can help us to better distinguish between samples from different classes.

Extreme right-wing violence has increased in recent years. Social media technologies and self-made media platforms have allowed extreme right-wing groups to create and push narratives online. These narratives have inspired others to act outside of the digital world and commit horrendous [GH1] attacks targeting minorities. These narratives are also published in a hybrid media system where audience-driven content is favored by social media, rather than objective and correct content. Therefore, it is important to understand the use of digital storytelling and social media technologies in a political context to push political views and stories online. This thesis studied how The Nordic Resistance Movement (NRM) uses digital storytelling strategies to build and push narratives on Twitter and self-made media platforms. This was conducted through a network and qualitative content analysis. 400 000 tweets were collected using Twitter’s API from 1/1-2022 to 31/3-2022. The outcome indicates that NRM dilutes its hateful narrative on Twitter but uses digital storytelling to evoke interest in the movement and combines it with influential hashtags according to the network analysis to force users onto self-made media platforms. This is when the narrative changes to the extreme. The study concludes that NRM uses social media in combination with its digital storytelling to attract users from traditional political spaces on Twitter through intriguing storytelling as part of an initiation process to attract new members and visitors to their self-made media platforms. Further research is suggested to be done on media ecology concerning self-made media platforms to gain further insight into how the platforms affect each other and the storytelling in the hybrid media system.

Idag är det vedertaget att det är vitalt för företag att involvera kunderna i produktutvecklingen för att bättre förstå marknadens behov och maximera värdeskapandet av produkterna. Denna involvering kommer att resultera i en ökad innovationskapacitet för organisationen. Denna fallstudie har för avsikt att besvara forskningsfrågan; hur ska ett medelstort högteknologiskt företag fördela kundinformationsflödet på ett systematiskt tillvägagångssätt för att öka innovationskapaciteten? Forskningsmetodiken består av tre delar; såsom förstudie, social nätverksanalys, samt intern och extern benchmarking. Förstudien består av 21 intervjuer internt i organisationen och det sociala nätverket baseras på en enkät, som 49 individer har besvarat. Benchmarkingen har involverat sex individer från en intern avdelning och två intervjuer med externa företag inom samma bransch. Resultatet tyder på att mängden interaktioner med externa kunder bör reduceras för att systematiskt och strukturerat inhämta kundinformationen. För att öka informationsflödet inom organisationen bör en särskild avdelning, som har daglig kontakt med alla avdelningar, ha en informationsspridande roll som överför informationen från marknadsavdelningen till resterande avdelningar internt. Därtill bör spridningen av kundinformation integreras i den dagliga arbetsprocessen, då det underlättar att anamma och använda informationen i det dagliga arbete. / Nowadays, it is well known that it is highly important to involve the customers in the product development in order to better understand the needs of the market, increase the relations to customers and maximize the value creation of the products. This will result in a higher innovation capacity for the organization. This case study intends to answer the research questions, how to allocate the customer information flow in a systematic approach at a medium size high technology company in order to increase the innovation capacity. The research methodology consists of three different parts including pre study, social network analysis, and internal and external benchmarking. The pre study consists of 21 interviews internally in the organization and the social network analysis is based on a survey, which 49 individuals have answered. The benchmarking involves six employees from another department and two interviews with external companies within the same business. The results indicate that the amount of customer interactions with external parties must be decreased in order to systematically receive and maintain customer information. In order to increase the information flow within the organization a specific department, that have daily contacts with almost all departments, should act as a transmitting function as they would connect marketing with research and development. Furthermore, the customer information should be included in the daily working process as it is easier for the employees to embrace and jointly utilize this information.

One important part of the management of zoo populations is the exchange of animals. The removal of an individual can have unknown effects on the social dynamics of the group. Social network studies are a well-established method to describe the social interactions within a group. This study aims to describe the social interactions in a group of chimpanzees and to test how social dominance rank predicts social interaction patters using a social network approach. Data was collected via observations on Kolmarden Wildlife Parks chimpanzee group, which is composed of seven males and eleven females. A total of 50 h of data was collected over 16 days. This resulted in a dominance rank and four different social networks for different behaviors (touch proximity, proximity, affiliative and agonistic behavior). The eigenvector coefficient, with the notable exception of the proximity network, was rarely correlated with the dominance rank and the highest ranked individual was never the most central. The more dominate individuals had fewer links to others through proximity and affiliative interactions. My analysis of the social network structure provides some evidence that the removal of high-ranking individuals would be unlikely to disproportionally affect the structure of the social network in this group. I highlight the possibility of further analysis such as knock-out analysis (where you examine the consequences of the removal of specific individuals) on existing data and argue that more observations would help to draw up a well-structured plan for translocations of individuals in this group. / En viktig del i förvaltningen av djurparkspopulationer är utbytet av djur. Att flytta en individ från en grupp kan ha okända effekter på gruppens sociala dynamik. Studier av djurs sociala nätverk är en väletablerad metod för att beskriva sociala interaktioner inom en grupp av djur. Syftet med denna studie är att beskriva de sociala interaktionerna i en grupp av schimpanser och testa huruvida en ranking av dominans förutspår mönster i dessa sociala interaktioner genom att tillämpa ’social network analysis’ metoden. Observationsdata samlades in på Kolmårdens djurparks schimpansgrupp, som består av sju hanar och elva honor. Totalt samlades 50 timmar av data under 16 dagar. Detta resulterade i en dominansrankning och fyra olika nätverk för olika typer av sociala interaktioner (närhet med beröring, närhet, affiliativa och agnostiska beteenden). Egienvector koefficienten, med det anmärkningsvärda undantaget för närhets nätverket, var sällan korrelerat med dominansrankningen och den högst rankade individen var aldrig mest central. Mer dominanta individer hade färre länkar till andra genom närhet med beröring och affiliativa interaktioner. Mina analyser av de sociala nätverkens struktur ger vissa bevis för att borttagandet av högt rankade individer inte skulle ge oproportionerliga effekter på den sociala strukturen i denna grupp. Jag uppmärksammar också möjligheterna att med mer analyser som t.ex. knock-out analyser (där man undersöker konsekvensen av att ta bort individer från olika nätverk) på befintlig data, samt mer observationer skulle hjälpa för att kunna göra en väl strukturerad plan för flytt av individer från denna grupp.

<p>Similarity in academic achievement among friends (i.e., academic homophily) can arise from two processes, selection and influence. This study applied stochastic actor-based modeling using SIENA to disentangle friendship selection and social influence regarding academic achievement of Chinese adolescents in a three-year longitudinal sample of 880 middle school students (400 girls, year 1 mean age = 13.33) and 525 high school students (284 girls, year 1 mean age = 16.45). SIENA analyses revealed significant selection and influence effects pertaining to academic achievement for both middle school students and high school students across three years. Chinese adolescents preferred friendships with similarly achieving or higher achieving peers but avoided friendships with lower achieving peers. Friends’ influence on academic achievement can be both beneficial and detrimental. Chinese adolescents were more likely to increase in achievement when they befriended high-achieving peers, but decreased achievement when they were friends with low-achieving peers. There were no significant sex differences or school year differences (i.e., the first to second years versus the second to third years) in selection or influence effects for academic achievement. Influence effects were stronger for middle school students than for high school students, but no significant grade level differences emerged for selection effects. This study expands upon prior research by simultaneously assessing selection and influence effects on academic achievement and further examining the direction and strength of selection and influence processes regarding academic achievement using sophisticated modeling analyses. These results provide insights into the important role of cultural context in peer relationships and academic development by considering the strong pressure Chinese adolescents experience to be academically successful.</p>

Motivation: Alzheimer’s Disease (AD) is a highly prevalent, neurodegenerative disease which causes gradual cognitive decline. As documented in the literature, evi- dence has recently mounted for the role of metabolic dysfunction in AD. Metabolomic data has therefore been increasingly used in AD studies. Metabolomic disease studies often suffer from small sample sizes and inflated false discovery rates. It is therefore of great importance to identify algorithms best suited for the inference of metabolic networks from small cohort disease studies. For future benchmarking, and for the development of new metabolic network inference methods, it is similarly important to identify appropriate performance measures for small sample sizes. Results: The performances of 13 different network inference algorithms, includ- ing correlation-based, regression-based, information theoretic, and hybrid methods, were assessed through benchmarking and structural network analyses. Benchmark- ing was performed on simulated data with known structures across six sample sizes using three different summative performance measures: area under the Receiver Op- erating Characteristic Curve, area under the Precision Recall Curve, and Matthews Correlation Coefficient. Structural analyses (commonly applied in disease studies), including betweenness, closeness, and eigenvector centrality were applied to simu- lated data. Differential network analysis was additionally applied to experimental AD data. Based on the performance measure benchmarking and network analysis results, I identified Probabilistic Context Likelihood Relatedness of Correlation with Biweight Midcorrelation (PCLRCb) (a novel variation of the PCLRC algorithm) to be best suited for the prediction of metabolic networks from small-cohort disease studies. Additionally, I identified Matthews Correlation Coefficient as the best mea- sure with which to evaluate the performance of metabolic network inference methods across small sample sizes. / Graduate

Protein molecules and their interactions via protein-protein interactions (PPIs) are at the core of cellular functions. While such global PPI networks have been useful for analyzing gene function and effects of genetic variants, they do not resolve tissue and cell-typespecific interactions. Here we leverage recent advances in single-cell RNA sequencing (scRNA-seq) to reconstruct cell-type-specific PPI networks across different tissues to enable a context-sensitive analysis of immune cells’ gene-protein pathways. Targeting B cells, T cells, and macrophage cells as a proof-of-principle, we used scRNA-seq data across different tissues from the Tabula Muris mouse consortium. We mapped the protein-coding DEGs to a protein-protein interaction network database (STRING v.11). Topological and global similarity analysis of the networks revealed distinct properties between tissues highlighting tissue-specific behaviors for each cell type. For example, we found that degree and clustering coefficients distributions were tissue-specific. Different cell types and tissues displayed specific characteristics, and in particular, the splenic PPI networks were different compared to other analyzed tissues for all the immune cell types examined. For example, the pairwise comparison of the Jaccard index for node similarity and the mantel test correlation analysis showed that the spleen’ node and PPI networks are more different than any other tissues for each cell type examined. The physiological and anatomical properties that distinguish the spleen from other examined tissues might explain why the splenic PPI networks tend to be less similar compared to other tissues. The cell-type-specific network analyses using the different distance measures between the adjacency matrices on the hub nodes such as Euclidean, Manhattan, Jaccard, and Hamming distances showed a macrophage-specific behavior not observed in B cells and T cells, confirming their lineage differences. Finally, we explored the rewiring of selected hub nodes and transcription factors in the PPI networks along with their biological enrichments to validate our observations. The suggested biological validity of our results confirms the relevance of data-driven reconstruction of these context-sensitive networks using more advanced network inference algorithms. In conclusion, scRNA-seq enables the reconstruction of global unspecific PPI networks into cell and tissue-specific networks, thereby providing an increased resolution of the biological context.

This thesis explores the interplay between digital social movements and political parties on social media. The overarching aim of the thesis is to contribute to the understanding of how the digitalization of the electoral arena has transformed social movements, and how this affects the political parties' perceptions of them. The theoretical framework consists of three analytical lenses to conceptualize this interplay, these are George and Leidner’s (2019) categorization and classification of digital activism, Gunnar Sjöblom’s (1968) theory on partystrategies in a multiparty system, and Anne Kaun’s (2017) concept desynchronization. Acombination of methods is used, including network analysis and interviews, to explore this interplay. The results of the study indicate that there is a desynchronization in the practices of digital social movements and Swedish political parties.