Omid Akbari
Papers
2
Total Citations
11
H-Index
2
About
Omid Akbari is a rising researcher at the intersection of fault-tolerant computing and energy-efficient hardware design, with a focus on ensuring the reliability of systems deployed in safety-critical environments. His primary research areas include approximate computing, fault tolerance, and the design of robust deep neural network (DNN) accelerators. Akbari’s major contributions center on developing novel frameworks that balance reliability with energy efficiency. His most cited work, "X-Rel," introduces an energy-efficient and low-overhead approximate reliability framework for error-tolerant applications in critical systems, achieving 9 citations since 2023. This work addresses the limitations of traditional Triple Modular Redundancy (TMR) by mitigating issues like design diversity and soft errors in nanoscale devices. More recently, in 2024, Akbari explored the use of the Posit number representation system to enhance fault tolerance in DNNs, a critical advancement for autonomous vehicles and robotics where a single hardware failure can have catastrophic consequences. By pioneering these low-overhead, energy-efficient solutions, Akbari is shaping the future of reliable computing for next-generation critical applications.
Research Focus
Key Achievements
Top Papers
- 1
- 2