Summary of Are Ai-generated Text Detectors Robust to Adversarial Perturbations?, by Guanhua Huang et al.
Are AI-Generated Text Detectors Robust to Adversarial Perturbations?
by Guanhua Huang, Yuchen Zhang, Zhe Li, Yongjian You, Mingze Wang, Zhouwang Yang
First submitted to arxiv on: 3 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper investigates the robustness of existing detectors for AI-generated text (AIGT) and proposes a novel detector, the Siamese Calibrated Reconstruction Network (SCRN), to address concerns about the potential misuse of large language models. The SCRN employs a reconstruction network to add and remove noise from text, extracting a semantic representation that is robust to local perturbations. The model also includes a siamese calibration technique to train the model to make equally confident predictions under different noise levels. Experiments on four publicly available datasets show that the SCRN outperforms existing methods by achieving 6.5% to 18.25% absolute accuracy improvement under adversarial attacks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary AI-generated text detectors lack robustness against small changes in characters or words, making it hard to distinguish between human-created and AI-generated text. The new Siamese Calibrated Reconstruction Network (SCRN) solves this problem by adding and removing noise from text, making the model more accurate and resistant to attacks. This helps prevent misuse of large language models. |
