Showing 1–12 of 12 results for author: Biswal, M

Z-AGI Labs at ClimateActivism 2024: Stance and Hate Event Detection on Social Media

Authors: Nikhil Narayan, Mrutyunjay Biswal

Abstract: In the digital realm, rich data serves as a crucial source of insights into the complexities of social, political, and economic landscapes. Addressing the growing need for high-quality information on events and the imperative to combat hate speech, this research led to the establishment of the Shared Task on Climate Activism Stance and Hate Event Detection at CASE 2024. Focused on climate activist… ▽ More In the digital realm, rich data serves as a crucial source of insights into the complexities of social, political, and economic landscapes. Addressing the growing need for high-quality information on events and the imperative to combat hate speech, this research led to the establishment of the Shared Task on Climate Activism Stance and Hate Event Detection at CASE 2024. Focused on climate activists contending with hate speech on social media, our study contributes to hate speech identification from tweets. Analyzing three sub-tasks - Hate Speech Detection (Sub-task A), Targets of Hate Speech Identification (Sub-task B), and Stance Detection (Sub-task C) - Team Z-AGI Labs evaluated various models, including LSTM, Xgboost, and LGBM based on Tf-Idf. Results unveiled intriguing variations, with Catboost excelling in Subtask-B (F1: 0.5604) and Subtask-C (F1: 0.7081), while LGBM emerged as the top-performing model for Subtask-A (F1: 0.8684). This research provides valuable insights into the suitability of classical machine learning models for climate hate speech and stance detection, aiding informed model selection for robust mechanisms. △ Less

Submitted 14 April, 2024; v1 submitted 26 February, 2024; originally announced February 2024.

Comments: Authors weren't supposed to upload given organisational agreements

Hate Speech and Offensive Content Detection in Indo-Aryan Languages: A Battle of LSTM and Transformers

Authors: Nikhil Narayan, Mrutyunjay Biswal, Pramod Goyal, Abhranta Panigrahi

Abstract: Social media platforms serve as accessible outlets for individuals to express their thoughts and experiences, resulting in an influx of user-generated data spanning all age groups. While these platforms enable free expression, they also present significant challenges, including the proliferation of hate speech and offensive content. Such objectionable language disrupts objective discourse and can… ▽ More Social media platforms serve as accessible outlets for individuals to express their thoughts and experiences, resulting in an influx of user-generated data spanning all age groups. While these platforms enable free expression, they also present significant challenges, including the proliferation of hate speech and offensive content. Such objectionable language disrupts objective discourse and can lead to radicalization of debates, ultimately threatening democratic values. Consequently, organizations have taken steps to monitor and curb abusive behavior, necessitating automated methods for identifying suspicious posts. This paper contributes to Hate Speech and Offensive Content Identification in English and Indo-Aryan Languages (HASOC) 2023 shared tasks track. We, team Z-AGI Labs, conduct a comprehensive comparative analysis of hate speech classification across five distinct languages: Bengali, Assamese, Bodo, Sinhala, and Gujarati. Our study encompasses a wide range of pre-trained models, including Bert variants, XLM-R, and LSTM models, to assess their performance in identifying hate speech across these languages. Results reveal intriguing variations in model performance. Notably, Bert Base Multilingual Cased emerges as a strong performer across languages, achieving an F1 score of 0.67027 for Bengali and 0.70525 for Assamese. At the same time, it significantly outperforms other models with an impressive F1 score of 0.83009 for Bodo. In Sinhala, XLM-R stands out with an F1 score of 0.83493, whereas for Gujarati, a custom LSTM-based model outshined with an F1 score of 0.76601. This study offers valuable insights into the suitability of various pre-trained models for hate speech detection in multilingual settings. By considering the nuances of each, our research contributes to an informed model selection for building robust hate speech detection systems. △ Less

Submitted 9 December, 2023; originally announced December 2023.

Comments: 14 pages, 3 figures. Accepted Working Notes at HASOC-FIRE 2023, to be published in CEUR Working Notes of FIRE

Analysis of Independent Learning in Network Agents: A Packet Forwarding Use Case

Authors: Abu Saleh Md Tayeen, Milan Biswal, Abderrahmen Mtibaa, Satyajayant Misra

Abstract: Multi-Agent Reinforcement Learning (MARL) is nowadays widely used to solve real-world and complex decisions in various domains. While MARL can be categorized into independent and cooperative approaches, we consider the independent approach as a simple, more scalable, and less costly method for large-scale distributed systems, such as network packet forwarding. In this paper, we quantitatively and… ▽ More Multi-Agent Reinforcement Learning (MARL) is nowadays widely used to solve real-world and complex decisions in various domains. While MARL can be categorized into independent and cooperative approaches, we consider the independent approach as a simple, more scalable, and less costly method for large-scale distributed systems, such as network packet forwarding. In this paper, we quantitatively and qualitatively assess the benefits of leveraging such independent agents learning approach, in particular IQL-based algorithm, for packet forwarding in computer networking, using the Named Data Networking (NDN) architecture as a driving example. We put multiple IQL-based forwarding strategies (IDQF) to the test and compare their performances against very basic forwarding schemes and simple topologies/traffic models to highlight major challenges and issues. We discuss the main issues related to the poor performance of IDQF and quantify the impact of these issues on isolation when training and testing the IDQF models under different model tuning parameters and network topologies/characteristics. △ Less

Submitted 4 February, 2022; originally announced February 2022.

Comments: 6 pages, 4 figures

AMI-FML: A Privacy-Preserving Federated Machine Learning Framework for AMI

Authors: Milan Biswal, Abu Saleh Md Tayeen, Satyajayant Misra

Abstract: Machine learning (ML) based smart meter data analytics is very promising for energy management and demand-response applications in the advanced metering infrastructure(AMI). A key challenge in developing distributed ML applications for AMI is to preserve user privacy while allowing active end-users participation. This paper addresses this challenge and proposes a privacy-preserving federated learn… ▽ More Machine learning (ML) based smart meter data analytics is very promising for energy management and demand-response applications in the advanced metering infrastructure(AMI). A key challenge in developing distributed ML applications for AMI is to preserve user privacy while allowing active end-users participation. This paper addresses this challenge and proposes a privacy-preserving federated learning framework for ML applications in the AMI. We consider each smart meter as a federated edge device hosting an ML application that exchanges information with a central aggregator or a data concentrator, periodically. Instead of transferring the raw data sensed by the smart meters, the ML model weights are transferred to the aggregator to preserve privacy. The aggregator processes these parameters to devise a robust ML model that can be substituted at each edge device. We also discuss strategies to enhance privacy and improve communication efficiency while sharing the ML model parameters, suited for relatively slow network connections in the AMI. We demonstrate the proposed framework on a use case federated ML (FML) application that improves short-term load forecasting (STLF). We use a long short-term memory(LSTM) recurrent neural network (RNN) model for STLF. In our architecture, we assume that there is an aggregator connected to a group of smart meters. The aggregator uses the learned model gradients received from the federated smart meters to generate an aggregate, robust RNN model which improves the forecasting accuracy for individual and aggregated STLF. Our results indicate that with FML, forecasting accuracy is increased while preserving the data privacy of the end-users. △ Less

Submitted 15 December, 2021; v1 submitted 12 September, 2021; originally announced September 2021.

Comments: 7 pages

$Z_N$ symmetry in $SU(N)$ gauge theories

Authors: Minati Biswal, Sanatan Digal, Vinod Mamale, Sabiar Shaikh

Abstract: We study $Z_N$ symmetry in $SU(N)$ gauge theories in the presence of matter fields in the fundamental representation, by restricting the lattice partition function integration to matter fields which are uniform in spatial directions and gauge fields with vanishing spatial components. In this approximation the gauge matter field interaction effectively reduces to a 1-dimensional gauged chain. This… ▽ More We study $Z_N$ symmetry in $SU(N)$ gauge theories in the presence of matter fields in the fundamental representation, by restricting the lattice partition function integration to matter fields which are uniform in spatial directions and gauge fields with vanishing spatial components. In this approximation the gauge matter field interaction effectively reduces to a 1-dimensional gauged chain. This makes analytical calculations of the matter field contribution to the Polyakov loop free energy possible. We show that in the limit of large number of temporal sites the explicit breaking of $Z_N$ symmetry in this free energy vanishes, driven by dominance of the density of states. We argue that the spatial links as well as the spatial modes of the matter fields determine the boundaries separating regions where $Z_N$ symmetry is realised from rest of the phase diagram. △ Less

Submitted 12 August, 2021; v1 submitted 25 February, 2021; originally announced February 2021.

Comments: 15 pages

Confinement-Deconfinement transition and $Z_2$ symmetry in $Z_2+$Higgs theory

Authors: Minati Biswal, Sanatan Digal, Vinod Mamale, Sabiar Shaikh

Abstract: We study the Polyakov loop and the $Z_2$ symmetry in the lattice $Z_2+$Higgs theory in 4D Euclidean space using Monte Carlo simulations. The results show that this symmetry is realised in the Higgs symmetric phase for large number of temporal lattice sites. To understand the dependence on the number of temporal sites, we consider a one dimensional model by keeping terms of the original action corr… ▽ More We study the Polyakov loop and the $Z_2$ symmetry in the lattice $Z_2+$Higgs theory in 4D Euclidean space using Monte Carlo simulations. The results show that this symmetry is realised in the Higgs symmetric phase for large number of temporal lattice sites. To understand the dependence on the number of temporal sites, we consider a one dimensional model by keeping terms of the original action corresponding to a single spatial site. In this approximation the partition function can be calculated exactly as a function of the Polyakov loop. The resulting free energy is found to have the $Z_2$ symmetry in the limit of large temporal sites. We argue that this is due to $Z_2$ invariance as well as dominance of the distribution or density of states corresponding to the action. △ Less

Submitted 17 July, 2021; v1 submitted 22 February, 2021; originally announced February 2021.

Comments: 17 pages, 23 figures

Re-visiting gravitational wave events via pulsars

Authors: Minati Biswal, Shreyansh S. Dave, Ajit M. Srivastava

Abstract: By now many gravitational wave (GW) signals have been detected by LIGO and Virgo, with the waves reaching earth directly from their respective sources. These waves will also travel to different pulsars and will cause (tiny) transient deformations in the pulsar shape. Some of us have recently shown that the resultant transient change in the pulsar moment of inertia may leave an observable imprint o… ▽ More By now many gravitational wave (GW) signals have been detected by LIGO and Virgo, with the waves reaching earth directly from their respective sources. These waves will also travel to different pulsars and will cause (tiny) transient deformations in the pulsar shape. Some of us have recently shown that the resultant transient change in the pulsar moment of inertia may leave an observable imprint on the pulsar signals as detected on earth, especially at resonance. The pulsars may thus act as remotely stationed Weber gravitational wave detectors. This allows us to revisit the past GW events via pulsars. We give here a list of specific pulsars whose future signals will carry the imprints of past GW events, to be specific we constrain it within 50 years. Some interesting cases are, supernova SN1987A with earliest perturbed signals from pulsars J0709-5923 and B0559-57 expected to reach earth in 2023 and 2024 respectively, Crab supernova, with perturbed signal arrival date from pulsar J1856-3754 in 2057, and GW170814 event with its imprints on the signals on the pulsar J0437-4715 reaching earth between 2035-2043. Even the earliest recorded supernova SN185 event may become observable again via pulsar J0900-3144 with the perturbed pulsar signal reaching us sometime between 2033-2066. Importantly, even though the strength of the signal will depend on the interior properties of the pulsar, the expected dates of signal arrival are completely model independent, depending only on the locations of the source and the relevant pulsar. △ Less

Submitted 10 September, 2019; originally announced September 2019.

Comments: 5 pages, 1 figure, 2 tables

Journal ref: Physical Letters B 811, (2020) 135887

$Z_3$ meta-stable states in PNJL model

Authors: Minati Biswal, Sanatan Digal, P. S. Saumia

Abstract: We study the Z 3 meta-stable states in the Polyakov loop Nambu-Jona-Lasinio (PNJL) model. These states exist for temperatures above T m ~ 194 MeV and can decay via bubble nucleation. We numerically solve the bounce equation to compute the nucleation rate. We speculate that, in the context of heavy-ion collisions, the likely scenario for the decay of the meta-stable states is via spinodal decomposi… ▽ More We study the Z 3 meta-stable states in the Polyakov loop Nambu-Jona-Lasinio (PNJL) model. These states exist for temperatures above T m ~ 194 MeV and can decay via bubble nucleation. We numerically solve the bounce equation to compute the nucleation rate. We speculate that, in the context of heavy-ion collisions, the likely scenario for the decay of the meta-stable states is via spinodal decomposition. △ Less

Submitted 5 October, 2020; v1 submitted 18 July, 2019; originally announced July 2019.

Comments: 24 pages, 11 figures (accepted by PRD)

Journal ref: Phys. Rev. D 102, 074020 (2020)

Chemically Induced Transformation of CVD-Grown Bilayer Graphene into Single Layer Diamond

Authors: Pavel V. Bakharev, Ming Huang, Manav Saxena, Suk Woo Lee, Se Hun Joo, Sung O Park, Jichen Dong, Dulce Camacho-Mojica, Sunghwan Jin, Youngwoo Kwon, Mandakini Biswal, Feng Ding, Sang Kyu Kwak, Zonghoon Lee, Rodney S. Ruoff

Abstract: Notwithstanding numerous density functional studies on the chemically induced transformation of multilayer graphene into a diamond-like film, a comprehensive convincing experimental proof of such a conversion is still lacking. We show that the fluorination of graphene sheets in Bernal (AB)-stacked bilayer graphene (AB-BLG) grown by chemical vapor deposition on a single crystal CuNi(111) surface tr… ▽ More Notwithstanding numerous density functional studies on the chemically induced transformation of multilayer graphene into a diamond-like film, a comprehensive convincing experimental proof of such a conversion is still lacking. We show that the fluorination of graphene sheets in Bernal (AB)-stacked bilayer graphene (AB-BLG) grown by chemical vapor deposition on a single crystal CuNi(111) surface triggers the formation of interlayer carbon-carbon bonds, resulting in a fluorinated diamond monolayer (F-diamane). Induced by fluorine chemisorption, the phase transition from AB-BLG to single layer diamond was studied and verified by X-ray photoelectron, ultraviolet photoelectron, Raman, UV-Vis, electron energy loss spectroscopies, transmission electron microscopy, and DFT calculations. △ Less

Submitted 7 January, 2019; originally announced January 2019.

Comments: 61 pages, 27 figures, 2 tables

Confinement-Deconfinement transition in $SU(2)+$Higgs Theory

Authors: Minati Biswal, Mridupawan Deka, Sanatan Digal, P. S. Saumia

Abstract: We study the confinement-deconfinement transition in $SU(2)$ gauge theory in the presence of massless bosons using lattice Monte Carlo simulations. The nature of this transition depends on the temporal extent ($N_τ$) of the Euclidean lattice. We find that the transition is a cross-over for $N_τ=2,4$ and second order with $3D$ Ising universality class for $N_τ=8$. Our results show that the second o… ▽ More We study the confinement-deconfinement transition in $SU(2)$ gauge theory in the presence of massless bosons using lattice Monte Carlo simulations. The nature of this transition depends on the temporal extent ($N_τ$) of the Euclidean lattice. We find that the transition is a cross-over for $N_τ=2,4$ and second order with $3D$ Ising universality class for $N_τ=8$. Our results show that the second order transition is accompanied by realization of the $Z_2$ symmetry. △ Less

Submitted 7 July, 2017; v1 submitted 26 October, 2016; originally announced October 2016.

Comments: 13 pages, 9 figures (Accepted in PRD)

Report number: IMSc/2016/10/06

Journal ref: Phys. Rev. D 96, 014503 (2017)

Dynamical Restoration of Z_N Symmetry in SU(N)+Higgs Theories

Authors: Minati Biswal, Sanatan Digal, P. S. Saumia

Abstract: We study the Z_N symmetry in SU(N)+Higgs theories with the Higgs field in the fundamental representation. The distributions of the Polyakov loop show that the Z_N symmetry is explicitly broken in the Higgs phase. On the other hand, inside the Higgs symmetric phase the Polyakov loop distributions and other physical observables exhibit the Z_N symmetry. This effective restoration of the Z_N symmetry… ▽ More We study the Z_N symmetry in SU(N)+Higgs theories with the Higgs field in the fundamental representation. The distributions of the Polyakov loop show that the Z_N symmetry is explicitly broken in the Higgs phase. On the other hand, inside the Higgs symmetric phase the Polyakov loop distributions and other physical observables exhibit the Z_N symmetry. This effective restoration of the Z_N symmetry changes the nature of the confinement-deconfinenement transition. We argue that the Z_N symmetry will lead to time independent topological defect solutions in the Higgs symmetric deconfined phase which will play important role at high temperatures. △ Less

Submitted 12 May, 2016; v1 submitted 26 November, 2015; originally announced November 2015.

Comments: 13 pages, 4 figures

NMR parameters in alkali, alkaline earth and rare earth fluorides from first principle calculations

Authors: Aymeric Sadoc, Monique Body, Christophe Legein, Mamata Biswal, Franck Fayon, Xavier Rocquefelte, Florent Boucher

Abstract: 19F isotropic chemical shifts for alkali, alkaline earth and rare earth of column 3 basic fluorides are measured and the corresponding isotropic chemical shieldings are calculated using the GIPAW method. When using PBE exchange correlation functional for the treatment of the cationic localized empty orbitals of Ca2+, Sc3+ (3d) and La3+ (4f), a correction is needed to accurately calculate 19F chemi… ▽ More 19F isotropic chemical shifts for alkali, alkaline earth and rare earth of column 3 basic fluorides are measured and the corresponding isotropic chemical shieldings are calculated using the GIPAW method. When using PBE exchange correlation functional for the treatment of the cationic localized empty orbitals of Ca2+, Sc3+ (3d) and La3+ (4f), a correction is needed to accurately calculate 19F chemical shieldings. We show that the correlation between experimental isotropic chemical shifts and calculated isotropic chemical shieldings established for the studied compounds allows to predict 19F NMR spectra of crystalline compounds with a relatively good accuracy. In addition, we experimentally determine the quadrupolar parameters of 25Mg in MgF2 and calculate the electric field gradient of 25Mg in MgF2 and 139La in LaF3 using both PAW and LAPW methods. The orientation of the EFG components in the crystallographic frame, provided by DFT calculations, is analysed in term of electron densities. It is shown that consideration of the quadrupolar charge deformation is essential for the analysis of slightly distorted environments or highly irregular polyhedra. △ Less

Submitted 17 May, 2011; v1 submitted 9 May, 2011; originally announced May 2011.

Comments: 18 pages, 8 figures, 4 tables and ESI