Hardware performance analysis: It
is often difficult to predict the performance of an application on a
new hardware without actually executing the application on the
hardware. My group works on developing analytical and empirical models
for application performance prediction on various hardware platforms.
[JETC 2024] [IISWC 2021]
Deployment of Edge-AI applications : My
group focuses on developing techniques for joint optimization of
accuracy- throughput-temperature trade-offs in practical deployments of
AI/ML applications (eg. traffic signal control) on Heterogenous MPSoCs.
[SEC 2023] [COMPASS 2022] [TCAD 2020]
Approximate Computing on Edge AI: We
focus on the design and implementation of techniques allowing AI models
and algorithms to be deployed on approximate hardware to achieve high
energy efficiency
without compromising the quality of results.