Papers
3
Total Citations
23
H-Index
2
About
Marcello Traiola is a researcher specializing in the reliability, fault tolerance, and hardware security of Deep Neural Networks (DNNs), with a particular focus on their deployment in safety-critical systems. His work addresses a crucial challenge at the intersection of machine learning and dependable computing: ensuring that neural networks remain accurate and trustworthy even when the underlying hardware is affected by faults or adversarial conditions. Traiola's most recognized contribution is *harDNNing* (2023, 12 citations), a machine-learning-based framework that systematically assesses and enhances the fault resilience of DNNs across demanding domains such as autonomous driving, aerospace, and smart healthcare. Complementing this, his 2021 work (9 citations) pioneered the integration of reliability and security mechanisms — particularly within Non-Volatile Memories — to simultaneously protect neural network intellectual property and strengthen fault resilience. His more recent 2024 study explores scalable reliability assessment methodologies for large DNN models, balancing performance trade-offs in safety-critical automotive and robotics applications. Through these contributions, Traiola has established himself as an important voice in making AI systems not only performant but genuinely dependable, bridging hardware engineering and modern deep learning in ways that matter profoundly for real-world deployment.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3