One pixel attack for fooling deep neural networks: Difference between revisions
Jump to navigation
Jump to search
| Line 15: | Line 15: | ||
We first formalize the generation of adversarial images as an constrained optimization problem. | We first formalize the generation of adversarial images as an constrained optimization problem. | ||
Let <math>f_t</math> | Let <math>\bold{x}<math> be the vectorized form of an image. Let <math>f_t(\bold{x})</math> | ||
= Evaluation and Results = | = Evaluation and Results = | ||
Revision as of 21:52, 25 March 2018
Presented by
1. Ziheng Chu
2. Minghao Lu
3. Qi Mai
4. Qici Tan
Introduction
Methodology
We first formalize the generation of adversarial images as an constrained optimization problem. Let [math]\displaystyle{ \bold{x}\lt math\gt be the vectorized form of an image. Let \lt math\gt f_t(\bold{x}) }[/math]