Search papers, labs, and topics across Lattice.
The paper introduces Explainable Authorship Variational Autoencoder (EAVAE) to disentangle style from content in text for improved authorship attribution and AI-generated text detection. EAVAE uses supervised contrastive learning to pretrain style encoders, followed by a VAE architecture with separate style and content encoders. A novel discriminator, which provides natural language explanations for its decisions, enforces disentanglement.
Achieve state-of-the-art authorship attribution and few-shot AI-generated text detection by explicitly disentangling style and content with a VAE architecture and an explanation-generating discriminator.
Learning robust representations of authorial style is crucial for authorship attribution and AI-generated text detection. However, existing methods often struggle with content-style entanglement, where models learn spurious correlations between authors'writing styles and topics, leading to poor generalization across domains. To address this challenge, we propose Explainable Authorship Variational Autoencoder (EAVAE), a novel framework that explicitly disentangles style from content through architectural separation-by-design. EAVAE first pretrains style encoders using supervised contrastive learning on diverse authorship data, then finetunes with a Variational Autoencoder (VEA) architecture using separate encoders for style and content representations. Disentanglement is enforced through a novel discriminator that not only distinguishes whether pairs of style/content representations belong to the same or different authors/content sources, but also generates natural language explanation for their decision, simultaneously mitigating confounding information and enhancing interpretability. Extensive experiments demonstrate the effectiveness of EAVAE. On authorship attribution, we achieve state-of-the-art performance on various datasets, including Amazon Reviews, PAN21, and HRS. For AI-generated text detection, EAVAE excels in few-shot learning over the M4 dataset. Code and data repositories are available online\footnote{https://github.com/hieum98/avae} \footnote{https://huggingface.co/collections/Hieuman/document-level-authorship-datasets}.
