References of "Postdoctoral thesis"
     in
Bookmark and Share    
Full Text
See detailLeveraging Graph Machine Learning for Social Network Analysis
Zhong, Zhiqiang UL

Postdoctoral thesis (2022)

As a ubiquitous complex system in quotidian life around everyone, online social networks (OSNs) provide a rich source of information about billions of users worldwide. To some extent, OSNs have mirrored ... [more ▼]

As a ubiquitous complex system in quotidian life around everyone, online social networks (OSNs) provide a rich source of information about billions of users worldwide. To some extent, OSNs have mirrored our real society: people perform a multitude of different activities in OSNs as they do in the offline world, such as establishing social relations, sharing life moments, and expressing opinions about various topics. Therefore, understanding OSNs is of immense importance. One key characteristic of human social behaviour in OSNs is their inter-relational nature, which can be represented as graphs. Due to sparsity and complex structure, analysing these graphs is quite challenging and expensive. Over the past several decades, many expert-designed approaches to graphs have been proposed with elegant theoretical properties and successfully addressed numerous practical problems. Nevertheless, most of them are either not data-driven or do not benefit from the rapidly growing scale of data. Recently, in the light of remarkable achievements of artificial intelligence, especially deep neural networks techniques, graph machine learning (GML) has emerged to provide us with novel perspectives to understanding and analysing graphs. However, the current efforts of GML are relatively immature and lack attention to specific scenarios and characteristics of OSNs. Based on the pros and cons of GML, this thesis discusses several aspects of how to build advanced approaches to better simplify and ameliorate OSN analytic tasks. Specifically: 1) Overcoming flat message-passing graph neural networks. One of the most widely pursued branches in GML research, graph neural networks (GNNs), follows a similar flat message-passing principle for representation learning. Precisely, information is iteratively passed between adjacent nodes along observed edges via non-linear transformation and aggregation functions. Its effectiveness has been widely proved; however, two limitations need to be tackled: (i) they are costly in encoding long-range information spanning the graph structure; (ii) they are failing to encode features in the high-order neighbourhood in the graphs as they only perform information aggregation across the observed edges in the original graph. To fill up the gap, we propose a novel hierarchical message-passing framework to facilitate the existing GNN mechanism. Following this idea, we design two practical implementations, i.e., HC-GNN and AdamGNN, to demonstrate the framework's superiority. 2) Extending graph machine learning to heterophilous graphs. The existing GML approaches implicitly hold a homophily assumption that nodes of the same class tend to be connected. However, previous expert studies have shown the enormous importance of addressing the heterophily scenario, where ``opposites attract'', is essential for network analysis and fairness study. We demonstrate the possibility of extending GML to heterophilous graphs by simplifying supervised node classification models on heterophilous graphs (CLP) and designing an unsupervised heterophilous graph representation learning model (Selene). 3) Online social network analysis with graph machine learning. As GML approaches have demonstrated significant effectiveness over general graph analytic tasks, we perform two practical OSN analysis projects to illustrate the possibility of employing GML in practice. Specifically, we propose a semantic image graph embedding (SiGraph) to improve OSN image recognition task with the associated hashtags semantics and a simple GNN-based neural link prediction framework (NeuLP) to boost the performance with tiny change. Keywords: Graph machine learning, Social network analysis, Graph neural networks, Hierarchical structure, Homophily/Heterophily graphs, Link prediction, Online image content understanding. [less ▲]

Detailed reference viewed: 98 (5 UL)
Full Text
See detailMotivation and self-regulation
Grund, Axel UL

Postdoctoral thesis (2019)

Detailed reference viewed: 85 (10 UL)
Full Text
See detailRational Architecture: Reasoning about Enterprise Dynamics
Van Zee, Marc UL

Postdoctoral thesis (2017)

Detailed reference viewed: 87 (22 UL)
See detailEurope's New Scientific Elite: Social Mechanisms of Science in the European Research Area.
Hoenig, Barbara UL

Postdoctoral thesis (2015)

The inquiry pursued the question whether the process of European integration in research funding has led to new forms of oligarchization and elite formation in the European Research Area. Empirically, we ... [more ▼]

The inquiry pursued the question whether the process of European integration in research funding has led to new forms of oligarchization and elite formation in the European Research Area. Empirically, we provided a case study on the European Research Council (ERC), which for the first time represents a case of institution-building in research funding at supranational level. The fact that the ERC intervenes into public science systems so far mostly organised at national level, thus simultaneously transforming traditional structures in the social organization of research, has motivated our hypothesis of an oligarchization process reproducing a new scientific elite in Europe. In scrutinising that presupposition we sought to identify those social mechanisms that generate, reproduce and possibly modify existing dynamics of stratification and oligarchization as an outcome of European integration in the area of science. [less ▲]

Detailed reference viewed: 379 (18 UL)
Full Text
See detailMeshless methods for shear-deformable beams and plates based on mixed weak forms
Hale, Jack UL

Postdoctoral thesis (2013)

Thin structural theories such as the shear-deformable Timoshenko beam and Reissner-Mindlin plate theories have seen wide use throughout engineering practice to simulate the response of structures with ... [more ▼]

Thin structural theories such as the shear-deformable Timoshenko beam and Reissner-Mindlin plate theories have seen wide use throughout engineering practice to simulate the response of structures with planar dimensions far larger than their thickness dimension. Meshless methods have been applied to construct numerical methods to solve the shear deformable theories. Similarly to the finite element method, meshless methods must be carefully designed to over- come the well-known shear-locking problem. Many successful treatments of shear-locking in the finite element literature are constructed through the application of a mixed weak form. In the mixed weak form the shear stresses are treated as an independent variational quantity in addition to the usual displacement variables. We introduce a novel hybrid meshless-finite element formulation for the Timoshenko beam problem that converges to the stable First-order/zero-order finite element method in the local limit when using maximum entropy meshless basis functions. The resulting formulation is free from the effects shear-locking. We then consider the Reissner-Mindlin plate problem. The shear stresses can be identified as a vector field belonging to the Sobelov space with square integrable rotation, suggesting the use of rotated Raviart-Thomas-Nedelec elements of lowest-order for discretising the shear stress field. This novel formulation is again free from the effects of shear-locking. Finally we consider the construction of a generalised displacement method where the shear stresses are eliminated prior to the solution of the final linear system of equations. We implement an existing technique in the literature for the Stokes problem called the nodal volume averaging technique. To ensure stability we split the shear energy between a part calculated using the displacement variables and the mixed variables resulting in a stabilised weak form. The method then satisfies the stability conditions resulting in a formulation that is free from the effects of shear-locking. [less ▲]

Detailed reference viewed: 481 (64 UL)