A Hybrid Deep BiLSTM-CNN for Hate Speech Detection in Multi-social media DOI Open Access
Ashwini Kumar, Santosh Kumar, Kalpdrum Passi

et al.

ACM Transactions on Asian and Low-Resource Language Information Processing, Journal Year: 2024, Volume and Issue: 23(8), P. 1 - 22

Published: May 6, 2024

Nowadays, means of communication among people have changed due to advancements in information technology and the rise online multi-social media. Many express their feelings, ideas, emotions on social media sites such as Instagram, Twitter, Gab, Reddit, Facebook, YouTube. However, misused send hateful messages specific individuals or groups create chaos. For various governance authorities, manually identifying hate speech platforms is a difficult task avoid In this study, hybrid deep-learning model, where bidirectional long short-term memory (BiLSTM) convolutional neural network (CNN) are used classify textual data, proposed. This model incorporates GLOVE-based word embedding approach, dropout, L2 regularization, global max pooling get impressive results. Further, proposed BiLSTM-CNN has been evaluated datasets achieve state-of-the-art performance that superior traditional existing machine learning methods terms accuracy, precision, recall, F1-score.

Language: Английский

Intelligent visual question answering in TCM education: An innovative application of IoT and multimodal fusion DOI
Wei Bi,

Qingyu Xiong,

Xingyi Chen

et al.

Alexandria Engineering Journal, Journal Year: 2025, Volume and Issue: 118, P. 325 - 336

Published: Jan. 23, 2025

Language: Английский

Citations

0
Figurative-cum-Commonsense Knowledge Infusion for Multimodal Mental Health Meme Classification DOI

Abdullah Mazhar,

Zuhair Hasan Shaik, Aseem Srivastava

et al.

Published: April 22, 2025

Language: Английский

Citations

0
DecodEM-X: advancing multimodal meme moderation with robust AI frameworks DOI
Hafiz Muhammad Arslan, Zhenhua Tan

Knowledge and Information Systems, Journal Year: 2025, Volume and Issue: unknown

Published: May 13, 2025

Language: Английский

Citations

0
Multimodal Hateful Meme Classification Based on Transfer Learning and a Cross-Mask Mechanism DOI Open Access
Fan Wu,

Chen Guolian,

Junkuo Cao

et al.

Electronics, Journal Year: 2024, Volume and Issue: 13(14), P. 2780 - 2780

Published: July 15, 2024

Hateful memes are malicious and biased sentiment information widely spread on the internet. Detecting hateful differs from traditional multimodal tasks because, in conventional tasks, visual textual align semantically. However, challenge detecting lies their unique nature, where images text may be weak or unrelated, requiring models to understand content perform reasoning. To address this issue, we introduce a fine-grained detection model named “TCAM”. The leverages advanced encoding techniques TweetEval CLIP introduces enhanced Cross-Attention Cross-Mask Mechanisms (CAM) feature fusion stage improve correlations. It effectively embeds features of data image descriptions into through transfer learning. This paper uses Area Under Receiver Operating Characteristic Curve (AUROC) as primary metric evaluate model’s discriminatory ability. approach achieved an AUROC score 0.8362 accuracy 0.764 Facebook Memes Challenge (FHMC) dataset, confirming its high capability. TCAM demonstrates relatively superior performance compared ensemble machine learning methods.

Language: Английский

Citations

3
Enhancing Multimodal Understanding With LIUS DOI Open Access

Chunlai Song

Journal of Organizational and End User Computing, Journal Year: 2024, Volume and Issue: 36(1), P. 1 - 17

Published: Jan. 12, 2024

VQA (visual question and answer) is the task of enabling a computer to generate accurate textual answers based on given images related questions. It integrates vision natural language processing requires model that able understand not only image content but also in order appropriate linguistic answers. However, current limitations cross-modal understanding often result models struggle accurately capture complex relationships between questions, leading inaccurate or ambiguous This research aims address this challenge through multifaceted approach combines strengths processing. By introducing innovative LIUS framework, specialized module was built process information fuse features using multiple scales. The insights gained from are integrated with “reasoning module” (LLM)

Language: Английский

Citations

0
Flexible margins and multiple samples learning to enhance lexical semantic similarity DOI
Jeng‐Shyang Pan, Xiao Wang, Dongqiang Yang

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 133, P. 108275 - 108275

Published: March 18, 2024

Language: Английский

Citations

0
A Hybrid Deep BiLSTM-CNN for Hate Speech Detection in Multi-social media DOI Open Access
Ashwini Kumar, Santosh Kumar, Kalpdrum Passi

et al.

ACM Transactions on Asian and Low-Resource Language Information Processing, Journal Year: 2024, Volume and Issue: 23(8), P. 1 - 22

Published: May 6, 2024

Nowadays, means of communication among people have changed due to advancements in information technology and the rise online multi-social media. Many express their feelings, ideas, emotions on social media sites such as Instagram, Twitter, Gab, Reddit, Facebook, YouTube. However, misused send hateful messages specific individuals or groups create chaos. For various governance authorities, manually identifying hate speech platforms is a difficult task avoid In this study, hybrid deep-learning model, where bidirectional long short-term memory (BiLSTM) convolutional neural network (CNN) are used classify textual data, proposed. This model incorporates GLOVE-based word embedding approach, dropout, L2 regularization, global max pooling get impressive results. Further, proposed BiLSTM-CNN has been evaluated datasets achieve state-of-the-art performance that superior traditional existing machine learning methods terms accuracy, precision, recall, F1-score.

Language: Английский

Citations

0