arXiv subjects

Vivek

Publications and source records attributed to Vivek.

4 recordsLinked to original sources

arXiv2025

Computational insights into Cobalt-based novel half-Heusler alloy for sustainable energy applications

The quest for efficient and sustainable green energy solutions has led to a growing interest in half Heusler alloys, particularly for thermoelectric and spintronic applications. This study investigates the multifaceted nature of cobalt based half Heusler alloy, CoVAs, employing DFT with advanced computational techniques, such as the FLAPW method. The elastic, electronic, magnetic, thermodynamic, and optical properties of CoVAs are meticulously analyzed. Structural and mechanical evaluations reveal mechanical stability and brittleness under varying pressures. Electronic and magnetic properties are examined through band structure and DOS analysis, revealing a half metallic nature with a minority spin band gap. The total magnetic moment aligns with the Slater Pauling rule, further confirming ferromagnetism and half metallicity. Thermodynamic investigations, based on the quasi-harmonic Debye approximation, provide insights into temperature- and pressure dependent behavior, including thermal expansion, heat capacity, and Debye temperature, establishing CoVAs as a viable candidate for high temperature applications. Additionally, the optical properties underestimate its potential in optoelectronic applications due to high absorption in the UV region, showing a distinct absorption edge corresponding to the electronic band gap. Phonon dispersion relations reflect the stability of the alloy, and the figure of merit confirms the alloy's suitability for thermodynamics applications. The findings highlight the potential of CoVAs as a promising candidate for spintronic photovoltaic and optoelectronic applications, providing insights into its fundamental properties that could facilitate experimental synthesis and industrial implementation for green energy and advanced technological applications.

cond-mat.mtrl-sci↗

arXiv2023

The Uli Dataset: An Exercise in Experience Led Annotation of oGBV

Online gender based violence has grown concomitantly with adoption of the internet and social media. Its effects are worse in the Global majority where many users use social media in languages other than English. The scale and volume of conversations on the internet has necessitated the need for automated detection of hate speech, and more specifically gendered abuse. There is, however, a lack of language specific and contextual data to build such automated tools. In this paper we present a dataset on gendered abuse in three languages- Hindi, Tamil and Indian English. The dataset comprises of tweets annotated along three questions pertaining to the experience of gender abuse, by experts who identify as women or a member of the LGBTQIA community in South Asia. Through this dataset we demonstrate a participatory approach to creating datasets that drive AI systems.

cs.CL↗

arXiv2023

Bulk and neutron-proton asymmetry coefficients of the semi-empirical mass formula tuned to ground state mass excess of AME2020 and/or FRDM(2012)

Davidson et al. has extended Seeger's mass formula to non-zero excitation energies by introducing temperature-dependent coefficients in the liquid drop energy part of the semi-empirical mass formula, without considering the shell effects. The semi-empirical mass formula of Davidson et al. is applicable for the compound nucleus temperatures less than or equal to 4 MeV. The mass excess calculated using this mass formula including shell effects/corrections does not reproduce the ground state mass excesses of the new atomic mass evaluation data AME2020 and/or FRDM(2012) with its coefficients at zero temperature. So, the coefficients of the semi-empirical mass formula are needed to be tuned to reproduce the ground state mass excess of the nuclei in the recent atomic mass evaluation data AME2020 and/or FRDM(2012). The bulk and neutron-proton asymmetry coefficients of the semi-empirical mass formula of Davidson et al. with shell effects have been tuned to reproduce the mass excess data for all the nuclei of AME2020 (Z=1-118 and A=1-295) and the nuclei of FRDM(2012) (Z=8-136 and A=16-339, except 3456 nuclei which are also available in the AME2020 data) at zero temperature, i.e., the coefficients are tuned for 9420 nuclei known at present. The tuned bulk and neutron-proton asymmetry coefficients reproduce the mass excess of the new atomic mass evaluation data AME2020 and/or FRDM(2012) within a difference of less than 1 MeV and can be used for the applications/investigations in the areas of physics where high energies are experienced or nuclei involved are in excited states, e.g., fusion-evaporation and fusion-fission processes in heavy-ion reactions.

nucl-th↗

arXiv2022

Robust Graph Neural Networks using Weighted Graph Laplacian

Graph neural network (GNN) is achieving remarkable performances in a variety of application domains. However, GNN is vulnerable to noise and adversarial attacks in input data. Making GNN robust against noises and adversarial attacks is an important problem. The existing defense methods for GNNs are computationally demanding and are not scalable. In this paper, we propose a generic framework for robustifying GNN known as Weighted Laplacian GNN (RWL-GNN). The method combines Weighted Graph Laplacian learning with the GNN implementation. The proposed method benefits from the positive semi-definiteness property of Laplacian matrix, feature smoothness, and latent features via formulating a unified optimization framework, which ensures the adversarial/noisy edges are discarded and connections in the graph are appropriately weighted. For demonstration, the experiments are conducted with Graph convolutional neural network(GCNN) architecture, however, the proposed framework is easily amenable to any existing GNN architecture. The simulation results with benchmark dataset establish the efficacy of the proposed method, both in accuracy and computational efficiency. Code can be accessed at https://github.com/Bharat-Runwal/RWL-GNN.

cs.LG↗