| KEYWORD |
Teaching Neuromorphic Hardware Fault Resiliency
keywords ARTIFICIAL INTELLIGENCE, HARDWARE ACCELERATORS, HARDWARE AND SOFTWARE, NEUROMORPHIC COMPUTING
Reference persons STEFANO DI CARLO, ALESSANDRO SAVINO
External reference persons Enrico Magliano
Research Groups DAUIN - GR-24 - SMILIES - reSilient coMputer archItectures and LIfE Sci
Thesis type RESEARCH / EXPERIMENTAL
Description Neuromorphic hardware, while powerful, can be prone to faults due to hardware degradation, environmental conditions, or manufacturing imperfections.
The human brain functions reliably, even in small damages or disruptions, offering inspiration for fault-resilient systems.
For neuromorphic hardware to be deployed in safety-critical applications (autonomous systems, healthcare devices, etc.), it must learn to adapt and
continue functioning even in the presence of faults.
Develop novel training methods that use fault injection to teach neuromorphic hardware how to detect, adapt, and recover from faults, ensuring long-term reliability and robustness.
Deadline 12/12/2025
PROPONI LA TUA CANDIDATURA