Robust Deep Learning

The Robust Deep Learning (RDL) research group, founded in 2023, is dedicated to advancing the development of deep learning models that are robust to various types of perturbations, including adversarial attacks, noisy data, and domain shifts. Our mission is to improve the reliability and stability of deep learning methods, enabling them to be applied effectively in real-world scenarios where data quality and integrity are often compromised.

Our research focuses on developing novel techniques for training deep neural networks that are resilient to various forms of perturbations. We explore strategies such as regularization, and data augmentation to enhance the robustness of deep learning models. Our group also has a particular interest in a dynamical systems perspective of neural networks, aiming to better understand the behavior of networks and motivate novel and more resilient architectures. Additionally, we investigate methods for detecting and mitigating adversarial and backdoor attacks.

The RDL group explores the application of robust deep learning in various domains, including computer vision, natural language processing, and science & engineering. We collaborate with industry partners and other research institutions to apply our methods to practical problems and evaluate their effectiveness in real-world settings. Overall, our group's goal is to push the boundaries of deep learning and improve its applicability in diverse settings by addressing the critical challenge of model robustness.

Our research has examined:

  • Robust sequence models
  • Robust transfer learning methods
  • Adversarially training¬†methods
  • Data augmentation methods
  • Backdoor detection methods
  • Continuous neural network architectures

The RDL Group is led by Dr. N. Benjamin Erichson, who also holds a joint appointment at the Lawrence Berkeley National Lab.


Principal Investigators and Research Scientists


  • Kareem Hegazy

Machine Learning Engineers

  • Jialin Song
  • Junyi Guo

Research Interns

  • Dongwei Lyu

Visiting Postdocs

  • Pu Ren

Visiting PhD Students

  • Ilan Naiman

Current Research Projects: