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.
Optimal Workplace Occupancy Strategies during the COVID-19 Pandemic
Authors:
Mansoor Davoodi,
Abhishek Senapati,
Adam Mertel,
Weronika Schlechte-Welnicz,
Justin M. Calabrese
Abstract:
During the COVID-19 pandemic, many organizations (e.g. businesses, companies, government facilities, etc.) have attempted to reduce infection risk by employing partial home office strategies. However, working from home can also reduce productivity for certain types of work and particular employees. Over the long term, many organizations therefore face a need to balance infection risk against produ…
▽ More
During the COVID-19 pandemic, many organizations (e.g. businesses, companies, government facilities, etc.) have attempted to reduce infection risk by employing partial home office strategies. However, working from home can also reduce productivity for certain types of work and particular employees. Over the long term, many organizations therefore face a need to balance infection risk against productivity. Motivated by this trade-off, we model this situation as a bi-objective optimization problem and propose a practical approach to find trade-off (Pareto optimal) solutions. We present a new probabilistic framework to compute the expected number of infected employees as a function of key parameters including: the incidence level in the neighborhood of the organization, the COVID-19 transmission rate, the number of employees, the percentage of vaccinated employees, the testing frequency, and the contact rate among employees. We implement the model and the optimization algorithm and perform several numerical experiments with different parameter settings. Furthermore, we provide an online application based on the models and algorithms developed in this paper, which can be used to identify the optimal workplace occupancy rate for real-world organizations.
△ Less
Submitted 4 April, 2022;
originally announced April 2022.
