Safe machine learning-based perception via closed-loop analysis

Placeholder Show Content


While machine learning enables new capabilities for the automation of safety-critical systems such as aircraft and cars, its introduction to these systems will require a significant validation effort. Machine learning presents new challenges in safe design and validation due to its complex nature. For this reason, the initial use of machine learning in these systems is likely to be limited to tasks that cannot be performed solely by traditional automation techniques. This thesis focuses on perceptual tasks that require processing high-dimensional image data from cameras to replace visual tasks typically performed by human operators. Specifically, we outline methods that rely on closed-loop analysis to encourage safety in machine learning-based perception systems. Our first two contributions relate to the safe design of neural network-based perception systems. Once designed, these systems should be put through a thorough verification process. The final two contributions of this thesis develop methods for the formal verification of these systems.


Type of resource text
Form electronic resource; remote; computer; online resource
Extent 1 online resource.
Place California
Place [Stanford, California]
Publisher [Stanford University]
Copyright date 2023; ©2023
Publication date 2023; 2023
Issuance monographic
Language English


Author Katz, Sydney Michelle
Degree supervisor Kochenderfer, Mykel J, 1980-
Thesis advisor Kochenderfer, Mykel J, 1980-
Thesis advisor Barrett, Clark
Thesis advisor Schwager, Mac
Degree committee member Barrett, Clark
Degree committee member Schwager, Mac
Associated with Stanford University, School of Engineering
Associated with Stanford University, Department of Aeronautics and Astronautics


Genre Theses
Genre Text

Bibliographic information

Statement of responsibility Sydney M. Katz.
Note Submitted to the Department of Aeronautics and Astronautics.
Thesis Thesis Ph.D. Stanford University 2023.

Access conditions

© 2023 by Sydney Michelle Katz
This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).

Also listed in

Loading usage metrics...