DINA: A Dual Defense Framework Against Internal Noise and External Attacks in Natural Language Processing

TL;DR

This paper introduces DINA, a unified framework that enhances NLP model robustness by simultaneously defending against internal label corruption and external adversarial attacks, crucial for safe AI deployment.

Contribution

DINA uniquely combines noisy-label learning and adversarial training to address dual threats in NLP, a novel approach not previously explored in this context.

Findings

DINA significantly improves robustness and accuracy over baseline models.

The framework effectively mitigates both internal and external adversarial threats.

Experimental results demonstrate practical applicability in real-world scenarios.

Abstract

As large language models (LLMs) and generative AI become increasingly integrated into customer service and moderation applications, adversarial threats emerge from both external manipulations and internal label corruption. In this work, we identify and systematically address these dual adversarial threats by introducing DINA (Dual Defense Against Internal Noise and Adversarial Attacks), a novel unified framework tailored specifically for NLP. Our approach adapts advanced noisy-label learning methods from computer vision and integrates them with adversarial training to simultaneously mitigate internal label sabotage and external adversarial perturbations. Extensive experiments conducted on a real-world dataset from an online gaming service demonstrate that DINA significantly improves model robustness and accuracy compared to baseline models. Our findings not only highlight the critical necessity of dual-threat defenses but also offer practical strategies for safeguarding NLP systems in realistic adversarial scenarios, underscoring broader implications for fair and responsible AI deployment.