Non-conventional Domain Research Articles

An Enhanced Filtering-Based Information Granulation Procedure for Graph Embedding and Classification

Granular Computing is a powerful information processing paradigm for synthesizing advanced pattern recognition systems in non-conventional domains. In this article, a novel procedure for the automatic synthesis of suitable information granules is proposed. The procedure leverages a joint sensitivity-vs-specificity score that accounts the meaningfulness of candidate information granules for each class considered in the classification problem at hand. Only statistically relevant granules are retained for a graph embedding procedure towards a geometric space, in which standard classification systems can be used without alterations. Performance tests have been carried out by considering open access datasets of fully labelled graphs with arbitrarily complex nodes and/or edges attributes that, by definition, must rely on inexact graph matching procedures to quantify dissimilarities. Two variants of the procedure are investigated: a standard variant, which aims at automatically finding suitable information granules for solving the classification problem as a whole, and a class-specific metric learning variant, in which the optimization procedure is performed in a class-aware fashion. In the latter case, each class will have its own set of information granules, along with the corresponding parameters defining distinct instances of the dissimilarity measure. Computational results show that the proposed algorithm is able to outperform the vast majority of current approaches for graph classification, while at the same time returning a grey-box model, interpretable by field-experts.

Finite elements with node dependent kinematics and scalable accuracy for the analysis of Stokes flows

This paper extends the use of one-dimensional elements with node-dependent kinematics to the analysis of Stokes flows. According to the Carrera Unified Formulation, the primary variables of the flow, velocity and pressure, are expressed as arbitrary expansions of the generalized unknowns. The novel implementation proposed in this work allows to increase the accuracy of the model only in the areas of the domain where it is required; i.e. close to boundaries, barriers or sudden expansion. Refined one-dimensional models based on Taylor and Lagrange expansions are used in this work and some typical applications are proposed to assess this novel technique, including Stokes flows in cylindrical and non-conventional domains. For each numerical example, different combinations of one-dimensional models have been considered to account for different kinematic approximations of flows, and the results, compared with analytical or finite volume solutions, highlight the capabilities of the proposed approach to handle non-conventional boundary conditions with ease and in preserving the computational cost without any accuracy loss.

Composite Eshelby model and domain band geometries of ferroelectric ceramics

A method of composite Eshelby inclusion is proposed for a ferroelectric grain with domain switching embedded in a polycrystalline ferroelectric matrix. The method quantifies the twinning structure due to spontaneous polarization, as well as the conventional and non-conventional domain structures after poling induced domain reorientation of 90 degree. The predicted parameters include the volume fraction, the thickness, and the surface inclination angle of switched domain plates. The domain wall energy for non-conventional domain structures is derived in terms of the arrays of misfit dislocations. The domain geometries predicted by the present work agree with the measured domain morphology near an indentation crack tip when subjected to lateral electric field.

REFLECTION IN OBJECT-ORIENTED PROGRAMMING

Recently, the notion of having programming languages and computational systems that allow their programs to reason about themselves and reflect on their computations has attracted the attention of many researchers. These systems, called "Reflective Systems", promise many advantages over conventional systems. For example, dynamic program stepping, execution tracing, and code analysis and modification at run time are few of the reflective operations that can be easily implemented in reflective systems. Reflectivity of programming languages allows the language users to extend the language to support other programming paradigms. Reflective systems can also play an important role in programming solutions for non-conventional domains. Artificial intelligence applications, particularly learning systems, benefit from the reflectivity of a programming language. A learning system must be able to modify itself at run time to incorporate its learned behaviors. The goal of this article is to define reflection and its components, introduce the appropriate terminology, and present briefly some reflective applications in object-oriented systems.