Summary of Trustworthy Actionable Perturbations, by Jesse Friedbaum et al.
Trustworthy Actionable Perturbations
by Jesse Friedbaum, Sudarshan Adiga, Ravi Tandon
First submitted to arxiv on: 18 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Information Theory (cs.IT)
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 This paper proposes a novel framework called Trustworthy Actionable Perturbations (TAP) that generates modified inputs, or counterfactuals, which change the true underlying class probabilities in a beneficial way. The TAP framework includes a verification procedure to ensure that these perturbations do not act as adversarial attacks and instead bring about desirable changes. To achieve this, the paper introduces new cost, reward, and goal definitions better suited for real-world applications. The authors also present PAC-learnability results for their verification procedure and theoretically analyze their method for measuring reward. In addition, they develop a methodology for creating TAP and compare its performance to previous counterfactual methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores ways to create modified inputs that change the true underlying probabilities in a beneficial way. It proposes a new framework called Trustworthy Actionable Perturbations (TAP) which includes a verification procedure to ensure the perturbations bring about desirable changes. The authors also introduce new cost, reward, and goal definitions that are better suited for real-world applications. |
