khoa16122004/RLAT-Reinforcement-Learing-in-Adversarial-Attack

This repo uses Deep Reinforcement-Learning Agent that can conduct a adversarial attack just using input and output of classification model.

Python

Black-box adversarial attack using Deeplearning

Problem statement

Can developers a agent can conduct a Black-box adversarial attack ?

Deep Reinforcement-Learning Agent

This agent using Deep Reinforcement-Learning Agent that can conduct a adversarial attack just using input and output of classification model.

Problem modeling

Alogorithm

DQN Traning Process