Overview
In 2021, I co-authored a research paper published at the CLEF 2021 CheckThat! Lab, targeting multiclass fake news detection using deep learning.
The system employs a Bidirectional LSTM (BiLSTM) architecture trained on a labelled dataset of news articles to classify them into multiple veracity categories. Text preprocessing, tokenisation, and embedding strategies were central to improving classification accuracy.
The paper was published in the CEUR Workshop Proceedings (Vol. 2936) and represents one of the earliest applications of BiLSTM to multiclass fake news classification in low-resource settings.
Key Highlights
- Published at CLEF 2021 CheckThat! Lab (CEUR-WS Vol. 2936)
- Multiclass fake news classification with BiLSTM
- Custom text preprocessing and embedding pipeline
- Benchmarked against baseline NLP classifiers