While our work lies at the
intersection of VLSI computer-aided design,
neuromorphic (brain-inspired) computing, circuit
architectures for hardware machine learning systems,
and applied machine learning, we currently focus on
the following three key areas:
Thrust
1: Neurally-Inspired Computing
Spike-dependent learning & training
algorithms
Energy-efficient hardware accelerator
architectures for spiking neural networks
Thrust
2: Machine Learning/Agentic-AI Enabled
Integrated Circuit Design and Test
Thrust
3: Applied Machine Learning
LLM-based Blackbox Optimization
Efficiency (e. g.,
sample-efficient semi-supervised
learning, compute/memory-efficient
generation)
Robustness (e.g., uncertainty
quantification of LLM generation,
hallucinations mitigation for
LLMs/VLMs)
