A project by RSim Research Group in collaboration with NVIDIA Research.
PyTorchFI is a runtime perturbation tool for deep neural networks (DNNs), implemented for the popular PyTorch deep learning platform. PyTorchFI enables users to perform perturbation on weights or neurons of a DNN during runtime. It is extremely versatile for dependability and reliability research, with applications including resiliency analysis of classification networks, resiliency analysis of object detection networks, analysis of models robust to adversarial attacks, training resilient models, and for DNN interpertability.
For example, this is an object detection network before a fault injection:
This is the same object detection network after a fault injection:
Our team added a tool within PyTorchFi capable of quickly developing graphs for your network injection. These graphs show where vulnerabilities lie within your neural network and the level of threat of these vulnerabilities.