Publication

Abstract :

Digital platforms have become one of the most powerful sources for spreading news and other information with much ease compared to traditional methods. However, this also made it easier to spread fake news and offensive content. A necessary strategy is required to detect fake news and offensive content to maintain the integrity of information and foster a safer online environment. The present study focuses on advancing the fake news and offensive content detection in Malayalam language, integrating transformer models with explainability techniques to render transparency and ensure high accuracy. We used multilingual transformer models such as mBERT, IndicBERT, XLM-RoBERTa, and MuRIL for classification tasks. Furthermore, comprehensive experiments with hyperparameter tuning are conducted on deep learning methods (CNN, LSTM, BiLSTM, GRU, BiGRU, CNN-LSTM, CNN-GRU) and traditional ML methods (SVM, LR, RF, DT, XGBoost, CATBoost, AdaBoost). The approach integrates TF-IDF and FastText embeddings for the representations, which is compared to BERT, IndicBERT, XLNet, mBERT, XLM-RoBERTa, and MuRIL. All models were evaluated using 10-fold cross-validation to ensure robustness and generalization of results. It is observed from the experiments that the XLM-RoBERTa model gave a better result with an F1 Score of 99.17% for the Fake news dataset and the mBERT model gave a better result with an F1 score of 95.95% for the offensive dataset. Furthermore, the present work utilizes explainable AI (XAI) approaches using LIME, Anchor, and Occlusion, to gain insights into the model decision making process. For the first time, an occlusion-based XAI method is proposed for fake news and offensive detection tasks. Also, the present work is the first in its kind to use Anchor and Occlusion based XAI for fake news and offensive content detection in Malayalam. The results showed that all three explainability methods occlusion, LIME and Anchor effectively captured the key words relevant for detecting both fake news and offensive content, with a high degree of overlap in the terms identified across models. This consistency demonstrates their strong adaptability and suitability for handling the complex linguistic and code-mixed characteristics of the Malayalam text.