A Unified Detection Framework for AI-Related Content and Artifacts
A unified detection framework for AI-related content using Mahalanobis distance scores is proposed, including methods for accurate positive class characterization and joint estimation.
Proposes a unified detection framework for various AI-related content using Mahalanobis distance scores and efficient estimators for positive class covariance matrices.
Keywords
Before reading this…
Applications
- →Detection of large language model generated text, hallucination, watermark, and adversarial examples
To understand this paper, make sure you know these concepts first:
- Understanding of artificial intelligence conceptsfind papers →
- Familiarity with statistical estimation methodsfind papers →
Abstract
More Like ThisArtificial intelligence (AI) is a double-edged sword: while it has achieved remarkable success across a wide range of domains, its deployment also calls for effective oversight and regulation, for which the detection of AI-related content and artifacts is perhaps the most direct and cost-effective approach. To this end, we propose a unified detection framework based on Mahalanobis distance scores (MDS), applicable to several important settings, including the detection of large language model (LLM) generated text, hallucination, watermark, and adversarial examples. A key component of the proposed method is to accurately characterize the positive class--such as human-generated text, factual statements, unwatermarked text, or non-adversarial samples--which requires an efficient and robust estimator of the covariance matrix of deep representations of positive samples before computing the MDS. Since the positive samples typically consist of multiple classes, and these classes may exhibit both homogeneity and heterogeneity, we develop joint estimation methods for both the casewise and cellwise minimum covariance determinant (MCD) estimators. We provide efficient optimization algorithms for both estimators and prove their convergence. We provide a reasonable definition of the breakdown point for the joint estimators and prove their corresponding high breakdown point properties. Empirical evaluations confirm the effectiveness of the proposed detection framework.