提出生长边界矩阵方法，提升NLP模型抗词语替换攻击能力。
Bridging Robustness and Generalization Against Word Substitution Attacks in NLP via the Growth Bound Matrix Approach
- 基于生长边界矩阵设计新型正则化，降低输入扰动影响
- 在多个数据集上提升对抗鲁棒性最高达8.8%
- 首次系统分析S4模型鲁棒性，适合关注安全NLP的研究者
尽管自然语言处理取得进展，模型仍易受同义词替换等对抗攻击影响。现有研究多聚焦前馈与卷积网络，对循环网络及现代状态空间模型（如S4）的鲁棒性研究不足，因其序列处理特性与复杂参数动态带来独特挑战。本文提出基于生长边界矩阵（GBM）的新正则化方法，旨在降低输入扰动对输出的影响。重点计算了LSTM、S4与CNN三种架构的GBM。该方法可（1）增强对词语替换攻击的抵抗能力，（2）提升干净文本上的泛化性能，（3）实现对状态空间模型（S4）鲁棒性的首次系统分析。在多个架构与基准数据集上的实验表明，本方法相比现有基线最多提升8.8%的对抗鲁棒性，优于多项先进防御方法。代码已开源。
原文摘要 · Abstract (English)
Despite advancements in Natural Language Processing (NLP), models remain vulnerable to adversarial attacks, such as synonym substitutions. While prior work has focused on improving robustness for feed-forward and convolutional architectures, the robustness of recurrent networks and modern state space models (SSMs), such as S4, remains understudied. These architectures pose unique challenges due to their sequential processing and complex parameter dynamics. In this paper, we introduce a novel regularization technique based on Growth Bound Matrices (GBM) to improve NLP model robustness by reducing the impact of input perturbations on model outputs. We focus on computing the GBM for three architectures: Long Short-Term Memory (LSTM), State Space models (S4), and Convolutional Neural Networks (CNN). Our method aims to (1) enhance resilience against word substitution attacks, (2) improve generalization on clean text, and (3) providing the first systematic analysis of SSM (S4) robustness. Extensive experiments across multiple architectures and benchmark datasets demonstrate that our method improves adversarial robustness by up to 8.8% over existing baselines. These results highlight the effectiveness of our approach, outperforming several state-of-the-art methods in adversarial defense. Codes are available at https://github.com/BouriMohammed/GBM
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。