Automated Fact Checking Using A Knowledge Graph-based Model

• Main Authors: Kundu, Arghya, Nguyen, Uyen Trang
• Format: Conference Proceeding
• Language: English
• Subjects: Convolutional neural networks; disinformation; Explainable AI; fact checking; fake news; Knowledge graphs; machine learning; Misinformation; natural language processing; Predictive models; Stability analysis; Training; Vectors
• ISSN: 2831-6983
• DOI: 10.1109/ICAIIC60209.2024.10463196

Misinformation is a growing threat to the economy, social stability, public health, democracy, and national security. One of the most effective methods to combat misinformation is fact checking. Fact checking is the process of verifying the factual accuracy of a statement or claim. Fact checkers employ rigorous methodologies to scrutinize claims, verify sources, and expose falsehoods. However, the huge volume of content circulating online makes it challenging for humans to identify misinformation manually. Automated tools can analyze large datasets to detect patterns in misinformation content, scaling up fact checking efforts. This paper proposes a knowledge graph-based fact checking model that uses two separate knowledge graphs, one containing true claims and the other, false claims. The model uses knowledge graph embeddings which are based on convolutional neural networks. The deep learning model is trained on the above two knowledge graphs to learn distinguishing patterns between true and false claims. Additionally, we employ explainable artificial intelligence (XAI) techniques to provide explanations for the model's classification, reducing cost of errors and increasing transparency and user trust in the system.