Research Projects

Compiler optimization helps programmers obtain better performance out of target hardware platforms. But besides performance optimization, compilers may have different objectives in compiling and optimizing a program. Some project topics:

• Compiler optimization for real-time systems: Optimize programs in a way that programs become more time predictable

• Compiler optimization for machine learning: Provide portable and efficient code generation for machine learning models

Real-time embedded systems range from autonomous cars to autonomous robots. Providing real-time guarantees in time-critical systems is essential for safety and reliability. Some project topics:

• Real-time dynamic neural networks: Dynamically change or update neural networks based on dynamic real-time requirements

• Real-time intermittent systems: Reflect energy availability in task scheduling for intermittent computing systems

Tiny machine learning is to enable machine learning applications on small embedded systems, which normally use low-power microcontrollers. Deploying machine learning models is challenging for small embedded systems because of limited resources. Some project topics:

• Tiny machine learning framework: Design a tiny machine learning framework for better resource management 

• Model compression: Design model compression and partitioning techniques for small embedded systems