Researchers have developed a new framework called Explainable Authorship Variational Autoencoder (EAVAE) to improve the accuracy and generalizability of authorship attribution and AI-generated text detection. This method disentangles writing style from content by using separate encoders and a novel discriminator that provides natural language explanations for its decisions. Experiments show EAVAE achieves state-of-the-art results on various authorship attribution datasets and performs well in few-shot learning for AI-generated text detection. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
RANK_REASON The submission is an arXiv preprint detailing a new method for authorship attribution and AI-generated text detection.
