AIRCC's International Journal of Computer Science and Information Technology

Open Access Open Access  Restricted Access Subscription Access

Multimodal Cyberbullying Meme Detection From Social Media Using Deep Learning Approach


Affiliations
1 Department of Information and Communication Technology, Comilla University, Bangladesh
2 Department of Electrical & Electronics Engineering, University of Rajshahi, Bangladesh
3 Department of Information and Communication Engineering, University of Rajshahi, Bangladesh
 

Cyberbullying includes the repeated and intentional use of digital technology to target another person with threats, harassment, or public humiliation. One of the techniques of Cyberbullying is sharing bullying memes on social media, which has increased enormously in recent years. Memes are images and texts overlapped and sometimes together they present concepts that become dubious if one of them is absent. Here, we propose a unified deep neural model for detecting bullying memes on social media. In the proposed unified architecture, VGG16-BiLSTM, consists of a VGG16 convolutional neural network for predicting the visual bullying content and a BiLSTM with one-dimensional convolution for predicting the textual bullying content. The meme is discretized by extracting the text from the image using OCR. The perceptron-based feature-level strategy for multimodal learning is used to dynamically combine the features of discrete modalities and output the final category as bullying or nonbullying type. We also create a “bullying Bengali memes dataset” for experimental evaluation. Our proposed model attained an accuracy of 87% with an F1 score of 88%. The proposed model demonstrates the capability to detect instances of Cyberbullying involving Bengali memes on various social media platforms. Consequently, it can be utilized to implement effective filtering mechanisms aimed at mitigating the prevalence of Cyberbullying.

Keywords

Cyberbullying, Multimodality, Bengali Meme, Deep Learning, NLP (Natural Language Processing), Machine Vision
User
Notifications
Font Size

  • M. A. Campbell, “Cyber Bullying An Old Problem in a New Guise?,” Aust. J. Guid. Couns., vol. 15, no. 1, pp. 68–76, 2005, doi: 10.1375/ajgc.15.1.68.

  • .Pawar and . . aje, “Multilingual cyberbullying detection system,” IEEE Int. Conf. Electro Inf. Technol., vol. 2019-May, pp. 040–044, 2019, doi: 10.1109/EIT.2019.8833846.

  • Velioglu and J. ose, “Detecting Hate peech in Memes sing Multimodal Deep Learning Approaches Prizewinning solution to Hateful Memes Challenge,” Dec. 2020, [Online]. Available: http://arxiv.org/abs/2012.12975.

  • S. Pramanick, . harma, D. Dimitrov, M. . Akhtar, P. Nakov, and T. Chakraborty, “MOMENTA A Multimodal Framework for Detecting Harmful Memes and Their Targets,” ep. 2021, [Online]. Available http://arxiv.org/abs/2109.05184.

  • T. Young, D. Hazarika, . Poria, and E. Cambria, “ ecent trends in deep learning based natural language processing [ eview Article],” IEEE Comput. Intell. Mag., vol. 13, no. 3, pp. 55–75, 2018, doi: 10.1109/MCI.2018.2840738.

  • A. Solanki, S. Kumar, and A. Nayyar, Emerging Trends and Applications of Machine Learning, vol. i. 2020.

  • M. Dadvar and K. Eckert, “Cyberbullying detection in social networks using deep learning based models,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 12393 LNCS, pp. 245–255, 2020, doi: 10.1007/978-3-030-59065-9_20.

  • F. Alam et al., “A eview of Bangla Natural Language Processing Tasks and the tility of Transformer Models,” 2021, [Online]. Available: http://arxiv.org/abs/2107.03844.

  • O. Sen et al., “Bangla natural language processing A comprehensive analysis of classical, machine learning, and deep learning-based methods,” IEEE Access, vol. 10, pp. 38999–39044, 2022, doi: 10.1109/ACCESS.2022.3165563.

  • M. . Karim, B. . Chakravarthi, J. P. McCrae, and M. Cochez, “Classification Benchmarks for nderresourced Bengali Language based on Multichannel Convolutional-L TM Network,” Apr. 2020, [Online]. Available http://arxiv.org/abs/2004.07807.

  • N. omim, M. Ahmed, M. . Islam, A. en harma, H. Talukder, and M. . Amin, “BD-SHS: A Benchmark Dataset for Learning to Detect Online Bangla Hate peech in Different ocial Contexts,” Proc. Lang. Resour. Eval. Conf., pp. 5153– 5162, 2022, [Online]. Available: https://aclanthology.org/2022.lrec-1.552.

  • M. T. Ahmed, M. ahman, . Nur, A. Z. M. T. Islam, and D. Das, “Natural language processing and machine learning based cyberbullying detection for Bangla and omanized Bangla texts,” Telkomnika (Telecommunication Comput. Electron. Control., vol. 20, no. 1, pp. 89–97, 2022, doi: 10.12928/TELKOMNIKA.v20i1.18630.

  • A. K. Das, A. Al Asif, A. Paul, and M. N. Hossain, “Bangla hate speech detection on social media using attention-based recurrent neural network,” J. Intell. Syst., vol. 30, no. 1, pp. 578–591, 2021, doi: 10.1515/jisys-2020-0060.

  • M. F. Ahmed, Z. Mahmud, Z. T. Biash, A. A. N. yen, and ..., “Cyberbullying detection using deep neural network from social media comments in bangla language,” arXiv Prepr. arXs 2021, [Online]. Available: https://arxiv.org/abs/2106.04506%0Ahttps://arxiv.org/pdf/2106.04506.

  • R.Ghosh, B. T. tudent, . Nowal, and G. Manju, “IJE T-Social Media Cyberbullying Detection using Machine Learning in Bengali Language ocial Media Cyberbullying Detection using Machine Learning in Bengali Language,” IJERT J. Int. J. Eng. Res. Technol., 2021, [Online]. Available: www.ijert.org.

  • M. R. Karim et al., DeepHateExplainer: Explainable Hate Speech Detection in Under-resourced Bengali Language, vol. 1, no. 1. Association for Computing Machinery, 2021. doi: 10.1109/DSAA53316.2021.9564230.

  • A. M. Ishmam and . harmin, “Hateful speech detection in public facebook pages for the bengali language,” Proc. - 18th IEEE Int. Conf. Mach. Learn. Appl. ICMLA 2019, no. December, pp. 555– 560, 2019, doi: 10.1109/ICMLA.2019.00104.

  • M. F. Mridha, M. A. H. Wadud, M. A. Hamid, M. M. Monowar, M. Abdullah-Al-Wadud, and A. Alamri, “L-Boost: Identifying Offensive Texts from Social Media Post in Bengali,” IEEE Access, vol. 9, pp. 164681–164699, 2021, doi: 10.1109/ACCESS.2021.3134154.

  • N. omim, M. Ahmed, H. Talukder, and M. aiful Islam, “Hate peech Detection in the Bengali Language A Dataset and Its Baseline Evaluation,” no. March 2022, pp. 457–468, 2021, doi: 10.1007/978-981-16-0586-4_37.

  • K. Kumari, J. P. ingh, Y. K. Dwivedi, and N. P. ana, “Towards Cyberbullying-free social media in smart cities: a unified multi-modal approach,” Soft Comput., vol. 24, no. 15, pp. 11059–11070, Aug. 2020, doi: 10.1007/s00500-01904550-x.

  • K. Kumari and J. P. ingh, “Identification of cyberbullying on multi-modal social media posts using genetic algorithm,” Trans. Emerg. Telecommun. Technol., vol. 32, no. 2, Feb. 2021, doi: 10.1002/ett.3907.

  • A. Kumar and N. achdeva, “Multimodal cyberbullying detection using capsule network with dynamic routing and deep convolutional neural network,” 2021. doi 10.1007/s00530-020-00747-5.

  • M. . Karim, . K. Dey, T. Islam, and B. . Chakravarthi, “Multimodal Hate Speech Detection from Bengali Memes and Texts,” Apr. 2022, [Online]. Available http://arxiv.org/abs/2204.10196.

  • K. melyakov, A. Chupryna, D. Darahan, and . Midina, “Effectiveness of modern text recognition solutions and tools for common data sources,” CEUR Workshop Proc., vol. 2870, pp. 154–165, 2021.

  • M. Ebersbach, . Herms, and M. Eibl, “Fusion methods for ICD10 code classification of death certificates in multilingual corpora,” CEUR Workshop Proc., vol. 1866, no. September, 2017.


Abstract Views: 248

PDF Views: 74




  • Multimodal Cyberbullying Meme Detection From Social Media Using Deep Learning Approach

Abstract Views: 248  |  PDF Views: 74

Authors

Md. Tofael Ahmed
Department of Information and Communication Technology, Comilla University, Bangladesh
Nahida Akter
Department of Information and Communication Technology, Comilla University, Bangladesh
Maqsudur Rahman
Department of Information and Communication Technology, Comilla University, Bangladesh
Abu Zafor Muhammad Touhidul Islam
Department of Electrical & Electronics Engineering, University of Rajshahi, Bangladesh
Dipankar Das
Department of Information and Communication Engineering, University of Rajshahi, Bangladesh
Md. Golam Rashed
Department of Information and Communication Engineering, University of Rajshahi, Bangladesh

Abstract


Cyberbullying includes the repeated and intentional use of digital technology to target another person with threats, harassment, or public humiliation. One of the techniques of Cyberbullying is sharing bullying memes on social media, which has increased enormously in recent years. Memes are images and texts overlapped and sometimes together they present concepts that become dubious if one of them is absent. Here, we propose a unified deep neural model for detecting bullying memes on social media. In the proposed unified architecture, VGG16-BiLSTM, consists of a VGG16 convolutional neural network for predicting the visual bullying content and a BiLSTM with one-dimensional convolution for predicting the textual bullying content. The meme is discretized by extracting the text from the image using OCR. The perceptron-based feature-level strategy for multimodal learning is used to dynamically combine the features of discrete modalities and output the final category as bullying or nonbullying type. We also create a “bullying Bengali memes dataset” for experimental evaluation. Our proposed model attained an accuracy of 87% with an F1 score of 88%. The proposed model demonstrates the capability to detect instances of Cyberbullying involving Bengali memes on various social media platforms. Consequently, it can be utilized to implement effective filtering mechanisms aimed at mitigating the prevalence of Cyberbullying.

Keywords


Cyberbullying, Multimodality, Bengali Meme, Deep Learning, NLP (Natural Language Processing), Machine Vision

References