[CVPR 2020] A Large-Scale Dataset for Real-World Face Forgery Detection

Implementation of Papers on Adversarial Examples

A Julia rewrite of Dynare: solving, simulating and estimating DSGE models.

Differentiable Optimizers with Perturbations in Pytorch

[CVPR 2018] Tensorflow implementation of NAG : Network for Adversary Generation

[ICLR'24] Official PyTorch Implementation of ContraLSP

Single-Cell (Perturbation) Model Library

Combinatorial prediction of therapeutic perturbations using causally-inspired neural networks

NPI Ephemeris Propagation Tool with Uncertainty Extrapolation

[ICML'24] Official PyTorch Implementation of TimeX++

Universal Adversarial Audio Perturbations

Repo of the paper "On the Robustness of Sparse Counterfactual Explanations to Adverse Perturbations"

Space Engineering 3 Course Work at University of Sydney

Building a multi-label classifier from scratch and using transfer learning for the PASCAL VOC image dataset.

Adversarial Attack using a DCGAN

Code to analyze high-density EEG and concurrent EMG/EOG datastreams during balance perturbations (replicates results from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6088363/)

Dark photon conversions in our inhomogeneous Universe. Code repository associated with the papers https://arxiv.org/abs/2002.05165 and https://arxiv.org/abs/2004.06733.

A deep convolutional neural network is used to explain the results of another one (VGG19).

DeepDefend is an open-source Python library for adversarial attacks and defenses in deep learning models, enhancing the security and robustness of AI systems.

In this work, we extend the FGSM method proposing multistep adversarial perturbation (MSAP) procedures to study the recommenders’ robustness under powerful methods. Letting fixed the perturbation magnitude, we illustrate that MSAP is much more harmful than FGSM in corrupting the recommendation performance of BPR-MF.