Research map / Adversarial robustness and security / Certified and empirical robustness
Towards Understanding the Dynamics of the First-Order Adversaries
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use projected gradient ascent for the inner adversary optimization with a sufficiently small learning rate | Two-layer neural networks with quadratic loss and small perturbation radius · Initial phase of adversarial training | iterations to reach an approximate local optimum: O(η^{-2}) | Helped |
| Use a smaller input scale, with the perturbation ratio fixed | Adversarial training · Initial phase | convergence speed | Helped |
| Use a smaller input scale, which reduces the perturbation radius at a fixed ratio | Adversarial loss landscape · Initial phase | landscape regularity | Helped |
| Use a larger input scale at a fixed perturbation ratio | Adversarial training · Initial phase | ease of escaping local maxima | Hurt |
Closest papers by meaning
- On the Loss Landscape of Adversarial Training: Identifying Challenges and How to Overcome Them (NeurIPS 2020 · similarity 0.90)
- Convergence of Adversarial Training in Overparametrized Neural Networks (NeurIPS 2019 · similarity 0.89)
- Over-parameterized Adversarial Training: An Analysis Overcoming the Curse of Dimensionality (NeurIPS 2020 · similarity 0.89)
- Clarifying the Behavior and the Difficulty of Adversarial Training (AAAI 2024 · similarity 0.89)
- Adversarial Training Should Be Cast as a Non-Zero-Sum Game (ICLR 2024 · similarity 0.88)
- If You Want to Be Robust, Be Wary of Initialization (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.