Reverse engineers a binary classifier using only 1 of the classes and without making queries.

Made for HackGT. This GAN model forges data similar to the one provided by NCR

reproduction of Thermometer Encoding: One Hot Way To Resist Adversarial Examples in pytorch

Undergraduate Thesis

Implementation of the NIPS paper: https://papers.nips.cc/paper/6908-adversarial-ranking-for-language-generation.pdf

Adversarial Machine Translation with pytorch

Guided Perturbations: Self-Corrective Behavior in Convolutional Neural Networks

Tensorflow Implementation of adversarial learning based adversarial example generator

This project focusses on creating an attacker against Deep Neural Networks making them misclassify an image. This is done to evaluate to robustness of the model.

Create certifiably secure scikit-learn compatible machine learning pipelines. Based on "Certified Defenses for Data Poisoning Attacks" by Steinhardt et al.

My entry for ICLR 2018 Reproducibility Challenge for paper Synthesizing robust adversarial examples https://openreview.net/pdf?id=BJDH5M-AW

Physical adversarial attack for fooling the Faster R-CNN object detector

Doing analysis on shared embedding space for the natural languages of English and Tamil

Implementation of the methods proposed in **Adversarial Training Methods for Semi-Supervised Text Classification** on IMDB dataset (without pre-training)

This tool can be used to find the most influential words on a document. We define most influential as the words that influence a trained classifier the most to give it a particular classification.

Implementation of the methods proposed in **Adversarial Training Methods for Semi-Supervised Text Classification** on IMDB dataset (without pre-training)

Projects for CS-839 Topics in Security (Spring 2018)

Proposed defenses against several adversarial attacks for speech to text systems

Adversarial Training for Neural Relation Extraction