Explainable Disentangled Representation Learning for Generalizable Authorship Attribution in the Era of Generative AI

TL;DR

This paper introduces EAVAE, a novel framework that disentangles style from content in text representations, improving authorship attribution and AI-generated text detection with enhanced interpretability and generalization.

Contribution

EAVAE is a new model that explicitly separates style and content representations using architectural design and a discriminator that provides explanations, leading to state-of-the-art results.

Findings

Achieves state-of-the-art accuracy on Amazon Reviews, PAN21, and HRS datasets.

Excels in few-shot AI-generated text detection on the M4 dataset.

Provides interpretable decisions through natural language explanations.

Abstract

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}.