Abstract: Due to the recent advances in machine learning techniques and

embedded systems hardware, autonomy has become a reachable goal for

many embedded systems domains. Unfortunately, it is not straightforward

to achieve autonomy in many safety-critical embedded systems that

require predictable timing correctness, one of the most important

tenets in certification required for such systems. An example is the

autonomous driving system, where timeliness of computations is an

essential requirement of correctness due to the interaction with the

physical world. In this talk, I will give an overview on our research

contributions on addressing several algorithmic and system-level

challenges towards ensuring timing-predictable autonomy, and a specific

illustration of our methodology on ensuring timing predictability while

simultaneously optimizing power and accuracy in a DNN-driven autonomous

embedded system.

Short Bio: Cong Liu is currently an associated professor in the Department of
Computer Science at the University of Texas at Dallas. His research
focuses on autonomous embedded systems and real-time systems. He has
received multiple best paper awards from top-tier venues including
RTAS’18, RTSS’09 & RTSS'17, and INFOCOM'17. He is a recipient of the
NSF CAREER Award in 2017.