Yiyu Shi
University of Notre Dame, IN·Engineering
Computer hardware raises two hard problems: how to fit intelligent, adaptive computing into tiny, power-starved devices, and how to make computing less wasteful for the planet. The projects here tackle both by designing algorithms and chip hardware together rather than separately.
- Fields
Engineering
Computer Science
Biomedical Engineering
- Topics
implantable medical devices
pacemakers
deep brain stimulation
- Grants
- 2 known grants, through Jul 2030
What they work on
- Develop lightweight state-space models, compact models of signals over time, that analyze physiological signals on implants with tight power and size budgets, using adaptive quantization to cut the bits each calculation needs.
- Build uncertainty-aware safety mechanisms using Bayesian confidence estimation, so an implant updates its model only when it is unsure and still acts safely.
- Design energy-efficient processing-in-memory hardware for implants, which computes where data is stored, and co-evolve it with learning algorithms using fault tolerance and reinforcement learning.
- Develop architecture and design-automation tools that map hardware designs onto REFRESH devices, built from retired FPGA chips joined by a 2.5D interposer.
- Model the carbon footprint of REFRESH FPGAs across fabrication, integration and packaging, and automatically pick connection layouts and chips based on how aged each chip is.
More about the research
One project targets medical implants that could learn and adjust on their own. The other reuses retired chips to cut the carbon cost of making new ones.
Why it matters
Patients with pacemakers, brain stimulators, cochlear implants and similar devices could receive care that adjusts to their changing bodies, with longer battery life and fewer tuning visits. The implants could also better serve people whose physiology differs from that assumed in standard device designs. Separately, reusing retired FPGA chips could reduce the carbon emitted by chip manufacturing and keep toxic electronic waste out of landfills.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Collaborative Research: CSR: Algorithm-Design-Technology Co-optimization for Next-Generation Self-evolving Implantable Devices
NSF · Aug 2026 – Jul 2030
Collaborative Research: DESC: Type II: REFRESH: Revisiting Expanding FPGA Real-estate for Environmentally Sustainability Heterogeneous-Systems
NSF · Oct 2023 – Sep 2027
Closest labs
Simon Peter
University of Washington, WA·Computer Science
Design a programmable memory controller that moves data among memory tiers connected by Compute Express Link based on usage patterns, module age and reliability.
Deliang Fan
Arizona State University Campus Immersion, AZ·Engineering·Early-career
Design, fabricate and validate a hybrid in-memory computing chip that combines RRAM, a non-volatile resistive memory, with SRAM to run deep neural networks with high energy efficiency.
Bo Yuan
Rutgers University-New Brunswick, NJ·Engineering·Early-career
Develop design principles for building domain expertise into learning-based robot motion planners, improving their reliability and performance in complex, resource-limited real-world settings.
The school
University of Notre Dame: 19 labs on PI Pages, $346M a year on research, 277 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.