Fernando Fernandes dos Santos
Université de Rennes, Institut de Recherche en Informatique et Systèmes Aléatoires
Papers
2
Total Citations
34
H-Index
2
About
Fernando Fernandes dos Santos is a leading researcher at the intersection of high-performance computing and hardware reliability, with a primary focus on ensuring the trustworthiness of artificial intelligence systems. His work critically addresses the vulnerability of Graphics Processing Units (GPUs) and Deep Neural Networks (DNNs) when deployed in safety-critical domains like autonomous vehicles, robotics, and aerospace. His most influential contribution, "A Multi-level Approach to Evaluate the Impact of GPU Permanent Faults on CNN's Reliability" (32 citations), provides a foundational methodology for understanding how hardware aging and faults degrade the accuracy of convolutional neural networks over a decade-long lifespan. More recently, his 2024 paper on "Reliability Assessment of Large DNN Models" tackles the essential trade-off between computational performance and fault tolerance in massive models. By developing frameworks that balance accuracy with robustness, Santos is enabling the safe deployment of AI in environments where failure is not an option. His work is pivotal for engineers and researchers designing resilient autonomous systems that must operate reliably for years under harsh conditions.
Research Focus
Key Achievements
Top Papers
- 1
- 2