Central Clustering of Attributed Graphs

Partitioning a data set of attributed graphs into clusters arises in different application areas of structural pattern recognition and computer vision. Despite its importance, graph clustering is currently an underdeveloped research area in machine learning due to the lack of theoretical analysis and the high computational cost of measuring structural proximities. To address the first issue, we introduce the concept of metric graph spaces that enables central (or center-based) clustering algorithms to be applied to the domain of attributed graphs. The key idea is to embed attributed graphs into Euclidean space without loss of structural information. In addressing the second issue of computational complexity, we propose a neural network solution of the K-means algorithm for structures (KMS). As a distinguishing feature to improve the computational time, the proposed algorithm classifies the data graphs according to the principle of elimination of competition where the input graph is assigned to the winning model of the competition. In experiments we investigate the behavior and performance of the neural KMS algorithm.     

Wireless community networks

An open standard IEEE 802.11 wireless local area networking technology provides an inexpensive and easy way to share bandwidth among multiple users. Such technology makes it possible to implement ad hoc wireless community networks using fixed wireless nodes rather than conventional base. A rural ad hoc network forwards packets from one wireless node to another to route data to a limited number of Internet gateways. A wireless community network uses two types of links between nodes. Point-to-multipoint links use omnidirectional antennas, while point-to-point links employ unidirectional antennas.     

Physical modeling of metalworking processes—I: Determination of large plastic strains

This article discusses the application of the visioplasticity method to the evaluation of large plastic strains such as those occurring in metalforming. Although this method can be used for any mode of deformation, its application to plane-strain deformation is treated here. The distortion of a quadrilateral element of a grid is tracked to compute strains during deformation. In each case any two lines of the quadrilateral can be used (length before and after deformation and direction cosines before deformation) to determine strains in the element. The method has been verified by application to basic cases of deformation such as uniaxial compression, tension, pure shear, and rotation of elements. The effect of choice of lines upon the results of strain calculation is also discussed.     

Supported titanium catalysts based on magnesium dichloride—silicon tetraethoxide adduct for propylene polymerization

Anhydrous magnesium dichloride was reacted with silicon tetraethoxide to form a solid adduct. The adduct was treated with titanium tetrachloride and an internal Lewis base to prepare supported titanium catalysts. Alkyl benzoate (PhCOOR, R = Me, Et, n - Bu) or dialkyl phthalate [Ph(COOR)₂, R = Me, Et, n - Bu] was used as an internal Lewis base. The prepared catalysts (MT) in combination with triethylaluminum as a cocatalyst polymerize propylene to yield polypropylene. The nature and concentration of internal Lewis base influence the composition, specific surface area, and performance of the catalysts. The addition of diphenyl dimethoxysilane into the MT/Et₃Al catalyst system increases the isotactic index of polypropylene from 75 to 97 at a Si/Al mol ratio of 0.05. The overall results indicate that a dialkyl phthalate-incorporated catalyst shows better physical, chemical, and polymerization characteristics compared to a corresponding alkyl benzoate catalyst. © 1997 John Wiley & Sons, Inc. J Appl Polym Sci 64: 1445–1450, 1997     

Kinetics of copolyurethane network formation

Cure kinetics for the formation of copolyurethane networks of various compositions based on hydroxy-terminated polybutadiene (HTPB), poly(12-hydroxy stearic acid-co-TMP) ester polyol(PEP), and different isocyanates has been studied through viscosity build up during the cure reaction. The viscosity(N)—time (t) plots conform to the equation N = ae^bt, where a and b are empirical constants, dependent on the composition and the nature of the polyols and the isocyanates. The rate constants (k) for viscosity build up, evaluated from the slopes of dN/dt versus N plots at different temperatures, were found to vary significantly from 0.0073 to 0.25 min⁻¹; and the activation energies for gelation were found to be in the range 20 to 40 kJ mol⁻¹. The results have been interpreted in terms of the dependence of the rate constants on structural characteristics of the prepolymers. © 1997 John Wiley & Sons, Inc. J Appl Polym Sci 66: 1795–1801, 1997     

Run-time data decomposition for parallel implementation of image processing and computer vision tasks

This paper presents several static and dynamic data decomposition techniques for parallel implementation of common computer vision algorithms. These techniques use the distribution of features in the input data as a measure of load for data decomposition. Experimental results are presented by implementing algorithms from a motion estimation system using these techniques on a hypercube multiprocessor. Normally in a vision system a sequence of algorithms is employed in which output of an algorithm is input to the next algorithm in the sequence. The distribution of features computed as a by-product of the current task is used to repartition the data for the next task in the system. This allows parallel computation of feature distribution, and therefore the overhead of estimating the load is kept small. It is observed that the communication overhead to repartition data using these run-time decomposition techniques is very small. It is shown that significant performance improvements over uniform-block-oriented partitioning schemes are obtained.     

Gut–Brain Axis as a Pathological and Therapeutic Target for Neurodegenerative Disorders

Human lifestyle and dietary behaviors contribute to disease onset and progression. Neurodegenerative diseases (NDDs), considered multifactorial disorders, have been associated with changes in the gut microbiome. NDDs display pathologies that alter brain functions with a tendency to worsen over time. NDDs are a worldwide health problem; in the US alone, 12 million Americans will suffer from NDDs by 2030. While etiology may vary, the gut microbiome serves as a key element underlying NDD development and prognosis. In particular, an inflammation-associated microbiome plagues NDDs. Conversely, sequestration of this inflammatory microbiome by a correction in the dysbiotic state of the gut may render therapeutic effects on NDDs. To this end, treatment with short-chain fatty acid-producing bacteria, the main metabolites responsible for maintaining gut homeostasis, ameliorates the inflammatory microbiome. This intimate pathological link between the gut and NDDs suggests that the gut-brain axis (GBA) acts as an underexplored area for developing therapies for NDDs. Traditionally, the classification of NDDs depends on their clinical presentation, mostly manifesting as extrapyramidal and pyramidal movement disorders, with neuropathological evaluation at autopsy as the gold standard for diagnosis. In this review, we highlight the evolving notion that GBA stands as an equally sensitive pathological marker of NDDs, particularly in Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis and chronic stroke. Additionally, GBA represents a potent therapeutic target for treating NDDs.     

Rotating anisotropic disc of uniform strength

A procedure to design a constant thickness composite disc of uniform strength by radially tailoring the anisotropic elastic constants is proposed. A special case of an isotropic disc with radially varying modulus is also examined. Analytical results are also compared with FEM calculations for two cases of radially varying anisotropy and for an isotropic disc with variable modulus.     

Study of matrix–filler interaction through correlations between structural and viscoelastic properties of carbonous-filler/polymer-matrix composites

Polymer matrix composites reinforced with carbonous fillers are of significant commercial importance thanks to their vast application base. As the performance of such composites largely depends on matrix–filler interaction, the present study is focused on the impact of surface chemical states of polymer matrix and carbonous filler on the viscoelastic performance of the composites. Here we report investigation of the filler–matrix interface through spectroscopic techniques such as X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy. Viscoelastic properties of various polymer matrix composites prepared by varying the filler volume fraction and/or the matrix/filler type have been studied through dynamic mechanical thermal analysis. Further, to understand the matrix–filler interaction, correlations between viscoelastic parameters and various structural parameters such as the surface area of filler and the surface chemical states of filler/matrix obtained through XPS have been studied. Strong correlations between the viscoelastic parameters and the matrix/filler surface chemical states have been observed, suggesting the XPS as an important tool to study the role of the surface functionalities present on the matrix/filler surface to define the matrix–filler interaction. The filler surface functionalities such as C bound O have been found more compatible with the polymers having aromatic ring in the repeat unit. © 2020 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2019 , 137, 48660.     

Grain size engineered Ba0.9Sr0.1Ti0.9Hf0.1O3-Na0.5Bi0.5TiO3 relaxor ceramics with improved energy storage performance

Novel lead-free diphasic (1-x)Ba_0.9Sr_0.1Ti_0.9Hf_0.1O₃-xNa_0.5Bi_0.5TiO₃ (BSTH-NBT) ceramic nanocomposites were synthesized via an economically viable modified mechano-chemical activation technique. In the present investigation, we have developed an energy storage composite material by systematically optimizing the charge transport behavior and charge storage characteristics between the ferroelectric BSTH and piezoelectric NBT phase. The composite with x = 0.09 NBT concentration has shown the best energy storage properties with 1.61 J/cm³ discharge energy density along with 80.1% energy efficiency. The BSTH and NBT had a synergetic effect on the ferroelectric properties of the composites. The improvement in ferroelectric and piezoelectric properties along with excellent aging characteristics in composite materials is mainly attributed to enhancement in microstructural density, grain boundary interface, and stress effects. The improved dispersibility and excellent compatibility between BSTH and NBT phase have resulted in approximately 20% enhancement in breakdown strength of composite compared to pure BSTH ceramic.