Summary of How Real Is Real? a Human Evaluation Framework For Unrestricted Adversarial Examples, by Dren Fazlija et al.
How Real Is Real? A Human Evaluation Framework for Unrestricted Adversarial Examples
by Dren Fazlija, Arkadij Orlov, Johanna Schrader, Monty-Maximilian Zühlke, Michael Rohs, Daniel Kudenko
First submitted to arxiv on: 19 Apr 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 proposed paper tackles the issue of creating more realistic and unpredictable adversarial examples in image-based machine learning models. These malicious inputs can mislead even state-of-the-art algorithms, potentially compromising safety-critical systems like autonomous vehicles. The study highlights the need for more human evaluation frameworks to assess the effectiveness of these attacks. To address this gap, the authors introduce SCOOTER, an evaluation framework that provides standardized guidelines and tools for conducting statistically significant human experiments. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper discusses a problem with machine learning models in images called “adversarial examples”. These are fake pictures that look normal to people but can trick computers into making wrong decisions. Before, researchers had to make these fake pictures slightly different from real ones so they would be noticeable if someone looked closely. But now, some experts think they can create fake pictures that are almost indistinguishable from real ones, which could cause even more problems for computer systems like self-driving cars. |
Keywords
» Artificial intelligence » Machine learning
