Faking Fake News for Real Fake News Detection: Propaganda-loaded Training Data Generation

Abstract

While there has been a lot of research and many recent advances in neural fake news detection, defending against human-written disinformation remains underexplored. Upon analyzing current approaches for fake news generation and human-crafted articles, we found that there is a gap between them, which can explain the poor performance on detecting human-written fake news for detectors trained on automatically generated data. To address this issue, we propose a novel framework for generating articles closer to human-written ones. Specifically, we perform self-critical sequence training with natural language inference to ensure the validity of the generated articles. We then explicitly incorporate propaganda techniques into the generated articles to mimic how humans craft fake news. Eventually, we create a fake news detection training dataset, PROPANEWS, which includes 2,256 examples. Our experimental results show that detectors trained on PROPANEWS are 7.3% to 12.0% more accurate for detecting humanwritten disinformation than for counterparts trained on data generated by state-of-the-art approaches.

Uncovering deep-rooted cultural differences (UNCOVER)

“…Researchers are thus looking for more advanced generative methods that can likewise subtly apply various propaganda techniques. Huang et al [6], developed an approach that introduces loaded language and appeal-to-authority techniques into legitimate articles. Their proposed methodology assigns an 'importance score' to every sentence in the original piece based on its relevance to a generated summary.…”

SWING: Balancing Coverage and Faithfulness for Dialogue Summarization

“…Factual Consistency Enhancement While factuality has been widely explored in the field of fact-checking and fake news detection (Thorne et al, 2018;Wadden et al, 2020;Huang et al, 2022b;Shu et al, 2018;Pan et al, 2021;Huang et al, 2022a), factual inconsistency remains a major challenge for abstractive summarization. One line of work attempts to improve the faithfulness of the generated summary with a separate correction model that corrects the errors made by the summarization model Cao et al, 2020;Fabbri et al, 2022b) or directly fix factual inconsistencies in the training data (Adams et al, 2022).…”

COVID-19 Fake News Detection With Pre-trained Transformer Models

“…Although old neural network models such as CNN and RNN are still used, pretrained transformer models have proven to be more efficient and accurate due to their improvements [17]. Transformer models started with the invention of BERT [13] [18] [19] [20] [21] [22] during 2018 [23], followed by its variations such as BART [24], ROBERTa [5] [18] [22] [25] [26], DeBERTa [18] [26] and Electra [21] [22] [27].…”