Comments (3)
Hi, thanks for your interest in our work. As you said, we adopted setting2
(black-box attack) where the attacker doesn't know the defense method. This is partially because IF-Defense is non-differentiable (hard to calculate the gradient of the optimization process), so we cannot adopt setting1
(white-box attach) against IF-Defense. Therefore, to perform a fair comparison, we test all the defense/attacks under setting2
.
As described above, it's hard to test IF-Defense under setting1
. Instead, I tried the so-called gray-box attack or adaptative attack, where the attacker has some knowledge of our defense method, and has special design to attack it. Please see our paper Section 4.2 Adaptive attack
part for the results. Overall, the Adaptive attack
indeed outperforms setting2 attack, but our defense accuracy is still okay.
from if-defense.
Btw, I do implement the drop-200
attack in this repo here, which can reproduce the results reported in the paper.
from if-defense.
Get it, thanks for your kind reply!
from if-defense.
Related Issues (10)
- Why the attack acc of attack_script and inference.py is different? HOT 4
- Save the generated attack images HOT 1
- Is there a training code for DUP-Net? HOT 1
- Could you please provide more details about training conv-o-net on modelnet40? HOT 5
- Could you send me the attack results of LG-GAN and AdvPC ? HOT 2
- Question about perturb attack on pointnet++ HOT 1
- Target label choosing stragegies HOT 2
- Question about the adaptive attack ? HOT 3
- the location of /pretrain folder HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from if-defense.