Put on your hacker hat 🎩 This Capture The Flag (CTF) challenge is your ticket to becoming an adversarial machine learning maestro!
Easy
Your mission, should you choose to accept it, is to craft a sentence so cunning that it:
- 😃 Sounds positive to human eyes
- 👾 But fools our AI into thinking it's negative!
- 🕵️♂️ Uncover the secrets of machine learning vulnerabilities Adversial Examples Attack
- 🧙♂️ Flex your creative muscles in AI security
- 🎓 Get hands-on with adversarial machine learning (no Ph.D. required!)
-
Clone this repo (it's like stealing the enemy's playbook):
git clone https://github.com/AI-Security-Research-Group/Adversarial-Examples-CTF.git cd Adversarial-Examples-CTF
-
Create a virtual environment (your secret base):
python3 -m venv venv source venv/bin/activate # On Windows: `venv\Scripts\activate`
-
Arm yourself with the required packages:
pip3 install -r requirements.txt
-
Launch the challenge:
python3 app.py
-
Navigate to
http://localhost:5000
and let the games begin! 🎮
- 👀 Scope out the battlefield: Check out the sample sentences
- 🧪 Craft your linguistic weapon: Create a positively negative sentence
- 🚀 Launch your attack: Hit that "Analyze Sentiment" button
- 🏆 Claim victory: Successfully fool the AI and grab that flag!
👀 Say "I won!🏆" on Discussion Forum if you solve the challange. (Do not post solutions!🚫)
When you outsmart the AI, you'll see a flag in this format:
AISRG-CTF{...}
That ...
is your proof of mastery. Wear it with pride! 🦚
Here's what you're up against:
And here's what victory looks like:
- 🌟 Star and watch our GitHub repo: AISRG
We're cooking up new challenges faster than an AI can say "I'm not a robot!" 🤖
Remember, in the world of AI security, the most powerful weapon is your creativity. Now go forth and may your adversarial examples be ever in your favor!
Happy Hacking, Future AI Tamer! 🦄🔐