Projects

This project designs certifiable defenses against data poisoning and backdoor attacks during training. High-quality, abundant data is crucial for training deep learning models to address complex problems. However, the integrity of this data is threatened by data poisoning attacks, where an attacker can subtly modify the training set to manipulate the model’s predictions.

This project proves and improves the robustness of deep learning models against inference-time adversarial examples. A unique characteristic of deep learning models is their vulnerability to malicious attacks, even when the underlying code implementations are correct.

This project ensures the correctness of deep learning training and inference codes built on top of TensorFlow/PyTorch. Deep learning models have made remarkable advancements in solving complex and diverse problems. The training and inference of these models are coded built on platforms like TensorFlow and PyTorch.