Enhancing Authorship Attribution through Embedding Fusion: A Novel Approach with Masked and Encoder-Decoder Language Models

TL;DR

This paper introduces a new embedding fusion framework using masked and encoder-decoder language models to improve the accuracy of distinguishing AI-generated text from human-written content, achieving over 96% accuracy.

Contribution

The paper presents a novel embedding fusion approach that combines multiple pre-trained language models to enhance authorship attribution accuracy for AI-generated texts.

Findings

Achieved over 96% classification accuracy.

Attained MCC greater than 0.93 on diverse datasets.

Demonstrated robustness across multiple LLM-generated texts.

Abstract

The increasing prevalence of AI-generated content alongside human-written text underscores the need for reliable discrimination methods. To address this challenge, we propose a novel framework with textual embeddings from Pre-trained Language Models (PLMs) to distinguish AI-generated and human-authored text. Our approach utilizes Embedding Fusion to integrate semantic information from multiple Language Models, harnessing their complementary strengths to enhance performance. Through extensive evaluation across publicly available diverse datasets, our proposed approach demonstrates strong performance, achieving classification accuracy greater than 96% and a Matthews Correlation Coefficient (MCC) greater than 0.93. This evaluation is conducted on a balanced dataset of texts generated from five well-known Large Language Models (LLMs), highlighting the effectiveness and robustness of our novel methodology.