# Addition of Projected Gradient Descent Attack Algorithm ### Task - Implementation and setting up of PGD Algorithm in the source code and its deployment on the main website. ### Difficulty Level : Intermediate ### Steps to be followed - - Add the respective attack function to [attack.py](https://github.com/dsgiitr/adversarial_lab/blob/master/attack.py) file. - The function should be of the following framework - ``` def pgd(parameters): model body return dict() ``` - Create the app-route python decorator for the algorithm in the [app.py](https://github.com/dsgiitr/adversarial_lab/blob/master/app.py) file. - The function should be of the following framework - ``` @app.route('/pgd', methods=['GET']) def pgd(): return render_template('pgd.html') ``` - Create the HTML template file (pgd.html) for the respective algorithm in the [template](https://github.com/dsgiitr/adversarial_lab/tree/master/template) folder ### Resources to be followed - [Relevant Paper : Towards Deep Learning Models Resistant to Adversarial Attacks](https://arxiv.org/pdf/1706.06083.pdf)
Addition of Projected Gradient Descent Attack Algorithm
Task - Implementation and setting up of PGD Algorithm in the source code and its deployment on the main website.
Difficulty Level : Intermediate
Steps to be followed -
Resources to be followed