Papers

5

Total Citations

25

H-Index

3

About

Robert Limas Sierra is an emerging researcher whose work sits at the critical intersection of hardware reliability, GPU architecture, and deep learning systems. His research focuses primarily on fault tolerance and resilience evaluation of Graphics Processing Units (GPUs), with particular emphasis on how hardware failures affect the execution of neural network workloads in safety-critical applications. Sierra's most significant contributions center on developing frameworks and methodologies for assessing the impact of permanent faults on GPUs running Deep Neural Networks (DNNs) and Convolutional Neural Networks (CNNs). His most-cited work (9 citations) investigates how hardware faults affect Tensor Core Units — specialized GPU accelerators fundamental to modern machine learning — exploring how different numerical representations influence structural resilience. Complementing this, he has developed systematic fault injection frameworks to evaluate neural network reliability under realistic hardware failure conditions. With a growing body of work accumulating over 25 citations since 2022, Sierra has also examined how GPU scheduling policies and large-scale DNN model configurations influence system reliability. His research carries direct relevance to automotive, aerospace, healthcare, and robotics industries, where deploying AI systems safely under hardware constraints remains an urgent and largely unsolved challenge.

Research Focus

Key Achievements

3
H-Index
5
Papers
25
Total Citations
5
Avg Citations/Paper
🏆 Most Cited Paper
Exploring Hardware Fault Impacts on Different Real Number Representations of the Structural Resilience of TCUs in GPUs
9 citations · 2024
📈 Most Prolific Year: 2024 (3 Papers)
🤝 Key Collaborators: 13
🏛 Institutions: Politecnico di Torino, Pedagogical and Technological University of Colombia

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago