Suchergebnisse

Online influence is the plinth of the social networks' effect in our lives and its impact has been steeply increasing. From viral marketing to political campaigns and from news adoption to disease transmission, the way we are influenced by others is more prevalent than ever. In this thesis, we address the problem of efficiently learning and analyzing influence representations for numerous graph mining problems that are apropos.The first half of the thesis is devoted to the problem of influence maximization, an NP-hard combinatorial optimization problem. The aim is to find the nodes in a network that can maximize the spread of information, where the spread is typically defined by random influence probabilities and simple diffusion models. To address this, in the thesis' first part, we devise a node representation learning model based on diffusion cascades along with an adaptation of a traditional influence maximization algorithm that utilizes the output of the model. This framework surpasses competitive methods, evaluated in terms of computational time and the influence of the predicted seeds in cascades of the immediate future.The second part is devoted to learning how to perform influence maximization. We develop a graph neural network that inherently parameterizes an upper bound of influence estimation, and train it on small simulated graphs. We experimentally show that it can provide accurate estimations faster than the alternatives for graphs 10 times larger than the train set. Furthermore, we use the models' predictions and representations to propose three new influence maximization methods. An adaptation of Cost Effective Lazy Forward that surpasses SOTA but with significant computational overhead, a Q-learning model that learns to retrieve seeds sequentially, and a submodular function that acts as proxy for the marginal gain and can be optimized adaptively and greedily with a theoretical guarantee. The latter strikes the best balance between efficiency and accuracy in our experiments.In the second half of ...

Online influence is the plinth of the social networks' effect in our lives and its impact has been steeply increasing. From viral marketing to political campaigns and from news adoption to disease transmission, the way we are influenced by others is more prevalent than ever. In this thesis, we address the problem of efficiently learning and analyzing influence representations for numerous graph mining problems that are apropos.The first half of the thesis is devoted to the problem of influence maximization, an NP-hard combinatorial optimization problem. The aim is to find the nodes in a network that can maximize the spread of information, where the spread is typically defined by random influence probabilities and simple diffusion models. To address this, in the thesis' first part, we devise a node representation learning model based on diffusion cascades along with an adaptation of a traditional influence maximization algorithm that utilizes the output of the model. This framework surpasses competitive methods, evaluated in terms of computational time and the influence of the predicted seeds in cascades of the immediate future.The second part is devoted to learning how to perform influence maximization. We develop a graph neural network that inherently parameterizes an upper bound of influence estimation, and train it on small simulated graphs. We experimentally show that it can provide accurate estimations faster than the alternatives for graphs 10 times larger than the train set. Furthermore, we use the models' predictions and representations to propose three new influence maximization methods. An adaptation of Cost Effective Lazy Forward that surpasses SOTA but with significant computational overhead, a Q-learning model that learns to retrieve seeds sequentially, and a submodular function that acts as proxy for the marginal gain and can be optimized adaptively and greedily with a theoretical guarantee. The latter strikes the best balance between efficiency and accuracy in our experiments.In the second half of ...

Online influence is the plinth of the social networks' effect in our lives and its impact has been steeply increasing. From viral marketing to political campaigns and from news adoption to disease transmission, the way we are influenced by others is more prevalent than ever. In this thesis, we address the problem of efficiently learning and analyzing influence representations for numerous graph mining problems that are apropos.The first half of the thesis is devoted to the problem of influence maximization, an NP-hard combinatorial optimization problem. The aim is to find the nodes in a network that can maximize the spread of information, where the spread is typically defined by random influence probabilities and simple diffusion models. To address this, in the thesis' first part, we devise a node representation learning model based on diffusion cascades along with an adaptation of a traditional influence maximization algorithm that utilizes the output of the model. This framework surpasses competitive methods, evaluated in terms of computational time and the influence of the predicted seeds in cascades of the immediate future.The second part is devoted to learning how to perform influence maximization. We develop a graph neural network that inherently parameterizes an upper bound of influence estimation, and train it on small simulated graphs. We experimentally show that it can provide accurate estimations faster than the alternatives for graphs 10 times larger than the train set. Furthermore, we use the models' predictions and representations to propose three new influence maximization methods. An adaptation of Cost Effective Lazy Forward that surpasses SOTA but with significant computational overhead, a Q-learning model that learns to retrieve seeds sequentially, and a submodular function that acts as proxy for the marginal gain and can be optimized adaptively and greedily with a theoretical guarantee. The latter strikes the best balance between efficiency and accuracy in our experiments.In the second half of ...

AbstractPalygorskite–TiO2 nanoparticles are frequently used as nanocatalysts. In the present study, two different nanocatalysts were developed based on the use of different palygorskite–TiO2 ratios: 40–60 and 10–90. The nanocomposites were investigated for the photocatalytic degradation of the common fungicide tebuconazole (TEB), under aquatic conditions. The samples were extensively characterized by X‐ray powder diffraction, attenuated total reflectance–Fourier transform infrared spectroscopy, scanning electron microscopy and N2 specific surface area (SSA) by Brunauer–Emmett–Teller (BET) analytical techniques. The TiO2 nanoparticles were successfully dispersed on the mineral's surfaces and the photocatalytic activity reached 88.4% for the palygorskite–TiO2 ratio of 40:60, where the dispersion was better as proved by the total pore volume and BET parameters (0.49 cm3/g and 258 m2/g compared to 0.33 cm3/g and 220 m2/g of the 10:90 ratio). The photocatalytic efficiency of the proposed materials was significantly higher than Degussa P25 (33.2%), and that makes the palygorskite–TiO2 nanocomposites very promising for advanced application in fungicides' degradation in aquatic environments.