[论文] [论文] Evaluating the Semantic-to-Geometric Gap in Adversarial Defe...

论文概要

研究领域: CV 作者: Christopher Burger, Christina Trotter, Joseph Carlisle, Charles Walter 发布时间: 2026-09-22 arXiv: 2609.26733

中文摘要

视觉语言模型（VLM）的快速发展对学术诚信构成系统性挑战：学生把图形化题目拍成单张图像提交即可绕过实质参与，我们称之为"轻率剽窃"。为给教育者提供关于 VLM 局限的可操作数据，我们研究启发式对抗图像变换——目标是在保持人类可解读的同时降低模型性能。两阶段评测入门课程评估：先人工评估 VLM 在电路图上的基线，再对拓扑结构（逻辑门）与坐标几何（卡诺图）做自动化大规模评测。发现：高能力 VLM 可表现出可观鲁棒性，但所有模型都受对抗扰动影响。结论：视觉扰动是可行的近期权宜之计，但长期评估安全要求教育者重新思考评估设计——VLM 性能仍在持续攀升。

原文摘要

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.

---

*自动采集于 2026-09-24*

#论文 #arXiv #CV #小凯