Users on social networks such as Twitter interact with each other without much knowledge of the real-identity behind the accounts they interact with. This anonymity has created a perfect environment for bot accounts to influence the network by mimicking real-user behaviour. Although not all bot accounts have malicious intent, identifying bot accounts in general is an important and difficult task. In the literature there are three distinct types of feature sets one could use for building machine learning models for classifying bot accounts. These feature-sets are: user profile metadata, natural language features (NLP) extracted from user tweets and finally features extracted from the the underlying social network. Profile metadata and NLP features are typically explored in detail in the bot-detection literature. At the same time less attention has been given to the predictive power of features that can be extracted from the underlying network structure. To fill this gap we explore and compare two classes of embedding algorithms that can be used to take advantage of information that network structure provides. The first class are classical embedding techniques, which focus on learning proximity information. The second class are structural embedding algorithms, which capture the local structure of node neighbourhood. We show that features created using structural embeddings have higher predictive power when it comes to bot detection. This supports the hypothesis that the local social network formed around bot accounts on Twitter contains valuable information that can be used to identify bot accounts.
An embedding is a mapping from a set of nodes of a network into a real vector space. Embeddings can have various aims like capturing the underlying graph topology and structure, node-to-node relationship, or other relevant information about the graph, its subgraphs or nodes themselves. A practical challenge with using embeddings is that there are many available variants to choose from. Selecting a small set of most promising embeddings from the long list of possible options for a given task is challenging and often requires domain expertise. Embeddings can be categorized into two main types: classical embeddings and structural embeddings. Classical embeddings focus on learning both local and global proximity of nodes, while structural embeddings learn information specifically about the local structure of nodes' neighbourhood. For classical node embeddings there exists a framework which helps data scientists to identify (in an unsupervised way) a few embeddings that are worth further investigation. Unfortunately, no such framework exists for structural embeddings. In this paper we propose a framework for unsupervised ranking of structural graph embeddings. The proposed framework, apart from assigning an aggregate quality score for a structural embedding, additionally gives a data scientist insights into properties of this embedding. It produces information which predefined node features the embedding learns, how well it learns them, and which dimensions in the embedded space represent the predefined node features. Using this information the user gets a level of explainability to an otherwise complex black-box embedding algorithm.
The elastic property of membranes self-assembled from AB diblock and ABA triblock copolymers, as coarse-grained model of lipids and the bolalipids, are studied using the self-consistent field theory (SCFT). Specifically, solutions of the SCFT equations, corresponding to membranes in different geometries (planar, cylindrical, spherical, and pore) have been obtained for a model system composed of amphiphilic AB diblock copolymers and ABA triblock copolymers dissolved in A homopolymers. The free energy of the membranes with different geometries is then used to extract the bending modulus, Gaussian modulus, and line tension of the membranes. The results reveal that the bending modulus of the triblock membrane is greater than that of the diblock membrane. Furthermore, the Gaussian modulus and line tension of the triblock membrane indicate that the triblock membranes have higher pore formation energy than that of the diblock membranes. The equilibrium bridging and looping fractions of the triblock copolymers are also obtained. Implications of the theoretical results on the elastic properties of biologically equivalent lipid bilayers and the bolalipid membranes are discussed.
Controlling the orientation of nanostructured thin films of block copolymers (BCPs) is essential for next-generation lithography using BCPs. According to conventional wisdom, the orientation of BCP thin films is mainly determined by molecular interactions (enthalpy-driven orientation). Here, we show that the entropic effect can be used to control the orientation of BCP thin films. Specifically, we used the architecture of star-block copolymers consisting of polystyrene (PS) and poly(dimethylsiloxane) (PDMS) blocks to regulate the entropic contribution to the self-assembled nanostructures. The study unequivocally demonstrate that for star-block copolymers with the same volume fractions of PS and PDMS, perpendicularly oriented BCP nanostructures could be induced via an entropic effect regulated by the number of arms. Also, the feasibility of using the star-block copolymer thin films for practical applications is demonstrated by using the thin film as a mask for nanolithography or as a template for the fabrication of nanoporous monolith.
We examine the behavior of lamellar phases of charged/neutral diblock copolymer thin films containing mobile ions in the presence of an external electric field. We employ self-consistent field theory and focus on the aligning effect of the electric field on the lamellae. Of particular interest are the effects of the mobile ions on the critical field, the value required to reorient the lamellae from the parallel configuration favored by the surface interaction to the perpendicular orientation favored by the field. We find that the critical field depends strongly on whether the neutral or charged species is favored by the substrates. In the case in which the neutral species is favored, the addition of charges decreases the critical electric field significantly. The effect is greater when the mobile ions are confined to the charged lamellae. In contrast, when the charged species is favored by the substrate, the addition of mobile ions stabilizes the parallel configuration and thus results in an increase in the critical electric field. The presence of ions in the system introduces a new mixed phase in addition to those reported previously.
The line tension or edge energy of bilayer membranes self-assembled from binary amphiphilic molecules is studied using self-consistent-field theory (SCFT). Specifically, solutions of the SCFT equations corresponding to an infinite membrane with a circular pore, or an open membrane, are obtained for a coarse-grained model in which the amphiphilic species and hydrophilic solvents are represented by ABandED diblock copolymers and C homopolymers, respectively. The edge energy of the membrane is extracted from the free energy of the open membranes. Results for membranes composed of mixtures of symmetric and cone- or inverse cone-shaped amphiphilic molecules with neutral and/or repulsive interactions are obtained and analyzed. It is observed that an increase in the concentration of the cone-shaped species leads to a decrease of the line tension. In contrast, adding inverse cone-shaped copolymers results in an increase of the line tension. Furthermore, the density profile of the copolymers reveals that the line tension is regulated by the distribution of the amphiphiles at the bilayer edge.
Blends of AB-diblock copolymers and C-homopolymers (AB/C), in which A and C are capable of forming hydrogen bonds, are used as a model system to examine and validate two approaches for modeling the effects of hydrogen bonding on the phase behavior of polymeric systems. The first model assumes a negative Flory–Huggins parameter between the A/C monomers. This attractive-interaction model reproduces qualitatively correct phase transition sequences as compared with experiments, but it fails to correctly describe the change of lamellar spacing induced by the addition of the C-homopolymers. The second model assumes that hydrogen bonding leads to A/C complexation. The interpolymer-complexation model predicts correctly the order–order phase transition sequences and the decrease of lamellar spacing at strong hydrogen bonding. The results demonstrate that hydrogen bonding of polymers should be modeled by the interpolymer-complexation model.