-
HCDIR: End-to-end Hate Context Detection, and Intensity Reduction model for online comments
Authors:
Neeraj Kumar Singh,
Koyel Ghosh,
Joy Mahapatra,
Utpal Garain,
Apurbalal Senapati
Abstract:
Warning: This paper contains examples of the language that some people may find offensive.
Detecting and reducing hateful, abusive, offensive comments is a critical and challenging task on social media. Moreover, few studies aim to mitigate the intensity of hate speech. While studies have shown that context-level semantics are crucial for detecting hateful comments, most of this research focuses…
▽ More
Warning: This paper contains examples of the language that some people may find offensive.
Detecting and reducing hateful, abusive, offensive comments is a critical and challenging task on social media. Moreover, few studies aim to mitigate the intensity of hate speech. While studies have shown that context-level semantics are crucial for detecting hateful comments, most of this research focuses on English due to the ample datasets available. In contrast, low-resource languages, like Indian languages, remain under-researched because of limited datasets. Contrary to hate speech detection, hate intensity reduction remains unexplored in high-resource and low-resource languages. In this paper, we propose a novel end-to-end model, HCDIR, for Hate Context Detection, and Hate Intensity Reduction in social media posts. First, we fine-tuned several pre-trained language models to detect hateful comments to ascertain the best-performing hateful comments detection model. Then, we identified the contextual hateful words. Identification of such hateful words is justified through the state-of-the-art explainable learning model, i.e., Integrated Gradient (IG). Lastly, the Masked Language Modeling (MLM) model has been employed to capture domain-specific nuances to reduce hate intensity. We masked the 50\% hateful words of the comments identified as hateful and predicted the alternative words for these masked terms to generate convincing sentences. An optimal replacement for the original hate comments from the feasible sentences is preferred. Extensive experiments have been conducted on several recent datasets using automatic metric-based evaluation (BERTScore) and thorough human evaluation. To enhance the faithfulness in human evaluation, we arranged a group of three human annotators with varied expertise.
△ Less
Submitted 20 December, 2023;
originally announced December 2023.
-
Pathologist-Like Explanations Unveiled: an Explainable Deep Learning System for White Blood Cell Classification
Authors:
Aditya Shankar Pal,
Debojyoti Biswas,
Joy Mahapatra,
Debasis Banerjee,
Prantar Chakrabarti,
Utpal Garain
Abstract:
White blood cells (WBCs) play a crucial role in safeguarding the human body against pathogens and foreign substances. Leveraging the abundance of WBC imaging data and the power of deep learning algorithms, automated WBC analysis has the potential for remarkable accuracy. However, the capability of deep learning models to explain their WBC classification remains largely unexplored. In this study, w…
▽ More
White blood cells (WBCs) play a crucial role in safeguarding the human body against pathogens and foreign substances. Leveraging the abundance of WBC imaging data and the power of deep learning algorithms, automated WBC analysis has the potential for remarkable accuracy. However, the capability of deep learning models to explain their WBC classification remains largely unexplored. In this study, we introduce HemaX, an explainable deep neural network-based model that produces pathologist-like explanations using five attributes: granularity, cytoplasm color, nucleus shape, size relative to red blood cells, and nucleus to cytoplasm ratio (N:C), along with cell classification, localization, and segmentation. HemaX is trained and evaluated on a novel dataset, LeukoX, comprising 467 blood smear images encompassing ten (10) WBC types. The proposed model achieves impressive results, with an average classification accuracy of 81.08% and a Jaccard index of 89.16% for cell localization. Additionally, HemaX performs well in generating the five explanations with a normalized mean square error of 0.0317 for N:C ratio and over 80% accuracy for the other four attributes. Comprehensive experiments comparing against multiple state-of-the-art models demonstrate that HemaX's classification accuracy remains unaffected by its ability to provide explanations. Moreover, empirical analyses and validation by expert hematologists confirm the faithfulness of explanations predicted by our proposed model.
△ Less
Submitted 20 October, 2023;
originally announced October 2023.
-
Scaling of reconnection parameters in magnetic island coalescence: Role of in-plane shear flow
Authors:
Jagannath Mahapatra,
Rajaraman Ganesh,
Abhijit Sen
Abstract:
A 2D incompressible viscoresistive-MHD model [Mahapatra et. al., Phys. Plasmas 28, 072103 (2021)] is used to study the scaling of reconnection parameters in magnetic island coalescence problem under two interesting scenarios. Firstly, the effect of changing island half-width at a fixed system size is investigated. As the island half-width increases, the total magnetic flux content increases result…
▽ More
A 2D incompressible viscoresistive-MHD model [Mahapatra et. al., Phys. Plasmas 28, 072103 (2021)] is used to study the scaling of reconnection parameters in magnetic island coalescence problem under two interesting scenarios. Firstly, the effect of changing island half-width at a fixed system size is investigated. As the island half-width increases, the total magnetic flux content increases resulting in an increase in upstream magnetic field, upstream velocity field and unnormalized reconnection rate while kee** the normalized rate, downstream magnetic field and current sheet length independent of it. Interestingly, the reconnection rate is found to be different from the upstream and downstream velocity as well as inverse aspect ratio of the current sheet, as opposed to the findings of the Sweet-Parker model. Secondly, the in-plane shear flow effects are studied, kee** the island width and system size fixed. Here thickness and length of the current sheet, the upstream magnetic and velocity field components, reconnection rate and time, current sheet inclination angle with shear flow length scale and amplitude are calculated. Interestingly, the inclination angle of the current sheet and the diffusion region are found to be different. These differences are more in stronger shear flows. These results are significantly different from the Harris sheet setup with shear flow. The results have potential application in the Earth magnetospheric observations and fusion plasma experiments.
△ Less
Submitted 9 November, 2022; v1 submitted 28 July, 2022;
originally announced July 2022.
-
Coupling of "cold" electron plasma wave via stationary ion inhomogeneity to the plasma bulk
Authors:
Sanjeev Kumar Pandey,
Jagannath Mahapatra,
Rajaraman Ganesh
Abstract:
Using high resolution kinetic (VPPM-OMP 1.0) and fluid (BOUT++) solvers, evolution of long-wavelength electron plasma wave (EPW) in the presence of stationary periodic ion background non-uniformity is investigated. Mode coupling dynamics between long-wavelength EPW mode of scale k and ion inhomogeneity of scale $k_{0}$ is illustrated. Validity of well known Bessel function $J_{n}(x)$ scaling in th…
▽ More
Using high resolution kinetic (VPPM-OMP 1.0) and fluid (BOUT++) solvers, evolution of long-wavelength electron plasma wave (EPW) in the presence of stationary periodic ion background non-uniformity is investigated. Mode coupling dynamics between long-wavelength EPW mode of scale k and ion inhomogeneity of scale $k_{0}$ is illustrated. Validity of well known Bessel function $J_{n}(x)$ scaling in the cold plasma approximation (i.e., when phase velocity $ω/k >> v_{thermal}$) along-with the effect of ion inhomogeneity amplitude (A) on temporal evolution of energy density in the long-wavelength EPW mode is investigated. Effect of finite system sizes on the Bessel $J_{n}(x)$ scaling is examined and scaling law for $τ_{FM}$ i.e the time required to attain first minimum of energy density of the corresponding perturbed mode (also called phase mixing time for $k \sim 0$ modes) versus ion inhomogeneity amplitude A obtained from both kinetic and fluid solutions for each of the cases studied, along-with some major differences in $τ_{FM}$ scaling for small system sizes is also reported.
△ Less
Submitted 15 June, 2022;
originally announced June 2022.
-
Effect of In-Plane Shear Flow on the Magnetic Island Coalescence Instability
Authors:
Jagannath Mahapatra,
Arkaprava Bokshi,
Rajaraman Ganesh,
Abhijit Sen
Abstract:
Using a 2D Viscoresistive Reduced MagnetoHydroDynamic (VR-RMHD) model, the magnetic island coalescence problem is studied in the presence of in-plane, parallel shear flows. Extending the analytical work of Waelbroeck et al [33] and Throumoulopoulos et al [34] in the sub-Alfvénic flow shear regime for Fadeev equilibrium, the super-Alfvénic regime is studied for the first time numerically. A wide ra…
▽ More
Using a 2D Viscoresistive Reduced MagnetoHydroDynamic (VR-RMHD) model, the magnetic island coalescence problem is studied in the presence of in-plane, parallel shear flows. Extending the analytical work of Waelbroeck et al [33] and Throumoulopoulos et al [34] in the sub-Alfvénic flow shear regime for Fadeev equilibrium, the super-Alfvénic regime is studied for the first time numerically. A wide range of values of shear flow amplitudes and shear scale lengths have been considered to understand the effect of sub-Alfvénic and super-Alfvénic flows on the coalescence instability and its nonlinear fate. We find that for flow shear length scales greater than the magnetic island size, the maximum reconnection rate decreases monotonically from sub-Alfvénic to super- Alfvénic flow speeds. For scale lengths smaller than the island size, the reconnection rate decreases upto a critical value $v_{0c}$, beyond which, the shear flow is found to destabilize the islands. The value of v0c decreases with decrease in the value of shear flow length scale. Interestingly, for our range of parameters, we find suppression of the Kelvin-Helmholtz instability in super-Alfvénic flows even when the shear scale length is smaller than the island width. Observation of velocity streamlines shows that the plasma circulation inside the islands has a stabilizing influence in strong shear flow cases. Plasma circulation is also found to be responsible for decrease in upstream velocity, causing less pile-up of magnetic flux on both sides of the reconnection sheet.
△ Less
Submitted 13 October, 2021; v1 submitted 2 February, 2021;
originally announced February 2021.
