Domain Gating Ensemble Networks for AI-Generated Text Detection
This addresses the critical need for robust AI text detection that adapts to new domains and models, offering a domain-adaptive solution for applications like content moderation and security.
The paper tackles the problem of detecting AI-generated text across unseen domains by introducing DoGEN, a technique that ensembles domain expert detectors using domain classifier weights, achieving state-of-the-art in-domain performance and outperforming larger models on out-of-domain detection.
As state-of-the-art language models continue to improve, the need for robust detection of machine-generated text becomes increasingly critical. However, current state-of-the-art machine text detectors struggle to adapt to new unseen domains and generative models. In this paper we present DoGEN (Domain Gating Ensemble Networks), a technique that allows detectors to adapt to unseen domains by ensembling a set of domain expert detector models using weights from a domain classifier. We test DoGEN on a wide variety of domains from leading benchmarks and find that it achieves state-of-the-art performance on in-domain detection while outperforming models twice its size on out-of-domain detection. We release our code and trained models to assist in future research in domain-adaptive AI detection.