949-824-9127

Mohammad Odema Received Best Paper Nomination at DATE 2023

Mohammad Odema Received Best Paper Nomination at Design Automation and Test in Europe (DATE) 2023 for project title: "HADAS:  Hardware-Aware Dynamic Neural Architecture Search for Edge Performance Scaling." Congrats Odema!

Mohsen Imani Won Young Faculty Award

Assistant Professor Mohsen Imani won the Young Faculty Award from Defense Advanced Research Projects Agency to recognize Prof. Imani as a “rising star” in research positions, supporting his work in […]

Trier Mortlock Won ARC Best Student Poster Award

Trier Mortlock won Best Student Poster Award from the Automotive Research Center (ARC) Program Review for project title: “Adaptive and Efficient Perception for Autonomous Ground Vehicles Operating in Highly Stochastic […]

Rozhin Yasaei Won 2nd Best Paper Award at WiCys 2023

Assistant Professor Rozhin Yasaei won 2nd Best Paper Award at Women in Cyber Security Conference 2023 with paper titled:  “Attack and Fault Detection in the Internet of Things using Graph […]

Quoc-Viet Dang Received Excellence in Digital Learning

Assistant Professor QV Dang received Excellence in Digital Learning from UCI Division of Teaching Excellence and Innovation to recognize his demonstrated excellence in undergraduate teaching at UCI. Congrats Prof. Dang!

Towards Finding the Optimization Potential of Fine-Grained Locks

Name:  Ahmad Nahian Chair: Professor Brian Demsky Date: November 30, 2023 Time:  9:30 AM Location: ISEB 1010 Committee: Professor Rainer Doemer, Professor Quoc-Viet Dang Title: Towards Finding the Optimization Potential […]

2023 MECPS Symposium

Event: MECPS Symposium 2023 Date: Friday, December 8, 2022 Time: 9:00am – 1:30pm Location: Interdisciplinary Science and Engineering Building (ISEB) View Details

Provably-Safe Offloading of Neural Network Controllers for Energy-Efficiency in Autonomous Driving Systems

DATE/TIME: Wednesday, January 23rd, 12:00pm
SPEAKER: Mohanad Odema
To mitigate the high energy demand of Neural Network (NN) based Autonomous Driving Systems (ADSs), we consider the problem of offloading NN controllers from the ADS to nearby edge-computing infrastructure, but in such a way that formal vehicle safety properties are guaranteed. In particular, we propose the EnergyShield framework, which repurposes a controller "shield" as a low-power runtime safety monitor for the ADS vehicle.