Certifiably Robust Neural ODE with Learning-Based Barrier Function

Runing Yang, Ruoxi Jia, Xiangyu Zhang, Ming Jin

Research output: Contribution to journalArticlepeer-review

2 Scopus Citations


Neural Ordinary Differential Equations (ODEs) have gained traction in many applications. While recent studies have focused on empirically increasing the robustness of neural ODEs against natural or adversarial attacks, certified robustness is still lacking. In this letter, we propose a framework for training a neural ODE using barrier functions and demonstrate improved robustness for classification problems. We further provide the first generalization guarantee of robustness against adversarial attacks using a wait-and-judge scenario approach.

Original languageAmerican English
Pages (from-to)1634-1639
Number of pages6
JournalIEEE Control Systems Letters
StatePublished - 2023

Bibliographical note

Publisher Copyright:
© 2017 IEEE.

NREL Publication Number

  • NREL/JA-2C00-84885


  • datadriven control
  • machine learning
  • Neural networks


Dive into the research topics of 'Certifiably Robust Neural ODE with Learning-Based Barrier Function'. Together they form a unique fingerprint.

Cite this