Research

Broad Interests

Heterogeneous computer systems and architectures

Systems software for run-time power/performance optimization

Embedded- and Edge- AI/ML

On-going Activities

Computational adaptivity in multi-modal machine learning

Resource efficient distributed DNN inference on collaborative heterogeneous edge clusters

Past Activities

On-line learning and self-awareness for run-time system software

Power/Performance optimization of Embedded Heterogeneous Multi-core Systems

Dynamic Approximation for Power/Performance Management

Resource Management of Many-core Systems