Adversarial Machine Learning in Image Recognition By: Tim Klem Advised by Graduate Student Vincent Bindschaedler
What is Adversarial Machine Learning?
Black-box Adversarial Attacks
Current Progress Built interface to interact with Face++ servers No API call caps! Explored image manipulation libraries in Python Created classification model to more closely model current literature
Future Goals Build locally-hosted machine learning model Reduce complexity of images Assess runtime of model training and inference