评估针对基于视觉-语言模型的抄袭行为的对抗性防御中的语义-几何差距
Evaluating the Semantic-to-Geometric Gap in Adversarial Defenses Against Vision-Language Model-Based Plagiarism
浏览论文内容
中文总结 AI 辅助
本研究评估了对抗性图像变换对VLM抄袭的防御效果，发现所有模型均易受扰动，视觉扰动仅作短期措施，长期需重新设计评估。
中文摘要 AI 辅助
视觉-语言模型（VLM）能力的快速提升对学术诚信构成了系统性挑战。VLM 现在允许学生通过将图形问题捕获并提交为单一图像来绕过有意义的参与，我们将这种行为定义为琐碎抄袭。为了向教育工作者提供关于 VLM 局限性的可操作数据，我们研究了启发式对抗性图像变换的有效性，这些变换旨在降低模型性能，同时保持人类可解释性。通过对入门评估的两阶段评估，我们手动评估了基线 VLM 在电路图上的性能，随后对拓扑结构（逻辑门）和坐标几何（卡诺图）进行了自动化大规模评估。我们发现，虽然高性能 VLM 可以表现出显著的鲁棒性，但所有模型都容易受到对抗性扰动的影响。我们得出结论，虽然视觉扰动作为近期可行的权宜之计，但鉴于 VLM 性能的持续提升，长期评估安全需要教育工作者重新审视评估设计。
英文摘要
The rapidly advancing capabilities of vision-language models (VLMs) present a systemic challenge to academic integrity. VLMs now allow students to bypass meaningful engagement by capturing and submitting graphical problems as singular images, a practice we define as trivial plagiarism. To provide educators with actionable data on VLM limitations, we investigate the efficacy of heuristic adversarial image transformations designed to degrade model performance while remaining human-interpretable. Through a two-phase evaluation of introductory assessments, we manually assess baseline VLM performance on circuit diagrams, followed by an automated large-scale evaluation of topological structures (logic gates) and coordinate geometry (Karnaugh maps). We find that while highly capable VLMs can exhibit appreciable robustness, all models suffer vulnerability to adversarial perturbations. We conclude that while visual perturbations act as a viable near-term stopgap, long-term assessment security requires educators to reapproach assessment design given continually increasing VLM performance.
发表机构
- Pelican Quantitative（鹈鹕量化）
- The University of Mississippi（密西西比大学）
机构由 AI 辅助整理，请以论文原文为准。
