| KEYWORD |
Area Engineering
Reliability of AI-based applications running on ARM architectures
Thesis in external company 
Thesis abroad
keywords AI RELIABITY, ARM ARCHITECTURE, IMAGE TEST LIBRARIES
Reference persons ANNACHIARA RUOSPO, EDGAR ERNESTO SANCHEZ SANCHEZ
External reference persons ARM Company tutors:
Marco Restifo, Luca di Mauro
Research Groups DAUIN - GR-05 - ELECTRONIC CAD and RELIABILITY GROUP - CAD
Description The project aims at studying the reliability of machine learning algorithms running on ARM architectures, and the possible implementation of detection and mitigation strategies. As a part of the state of the art, a possible solution could be created by resorting to the use of an image test library.
On these regards, the proposal consists on:
• Developing a image test library for a machine learning algorithm on Arm architecture (Cortex M-TBD)
• developing strategies to assess and quantifying the reliability of a machine learning algorithms at system, software, and hardware levels on Arm architectures,
• developing strategies to hardening machine learning algorithms at system, software, and hardware levels on Arm architectures,
• deploying the strategies for quantifying and hardening the machine learning reliability on Arm architectures.
Required skills The candidate must count with a good knowledge and familiarity with programming languages such as C and Python. In addition, she or he is also required to have good experience with the use of hardware description languages, as well as good knowledge on hardware simulation, validation and verification. Even if not strictly required, the candidate may count with some experience with artificial intelligence applications for embedded systems based on Neural Networks.
Notes The thesis is to be developed in the Sophia Antipolis quarters of the partner company under the development of an internship.
Deadline 14/06/2025
PROPONI LA TUA CANDIDATURA