VP (Virtual Probe)
 |
We exploit
advanced statistical methods to model the spatial variation of
integrated circuits. Our goal is to minimize the testing and
characterization cost and, hence, facilitate adaptive tuning for
high-performance integrated circuits. This project involves various
aspects of semiconductor manufacturing, regression modeling and
statistical inference. |
BCI (Brain Computer Interface)
 |
We work
with UPMC (University of Pittsburgh Medical Center) to analyze
electrical and magnetic signals from human brain. Our goal is to help
the patient to re-gain a number of physical activities, even if his/her
neuron system is partially damaged. For example, a patient can control
a computer just by "thinking", instead of using a keyboard/mouse. This
project involves various aspects of signal processing, machine learning
and hardware implementation. |
IMPACT (Improving Chip Design via Computational Thinking)
 |
We apply
statistical machine
learning to model, analyze and optimize large-scale integrated
circuits. Our goal is to maximize circuit performance, improve circuit
robustness and reduce manufacturing cost. This project involves various
aspects of circuit design, regression modeling, yield prediction and
stochastic optimization. |