Hui Guan
University of Massachusetts, MA·Computer Science·Early-career
Deep learning powers voice assistants, cameras that recognize objects, and robots, but it usually needs powerful computers. Many applications would work better if the computing happened on small, nearby devices such as edge servers, sensors or robots.
- Fields
Computer Science
Engineering
- Topics
edge computing
edge intelligence
deep learning systems
- Grants
- 2 known grants, through Feb 2029
What they work on
- Build deep learning systems that adjust in real time which deep neural networks run, responding to task difficulty, changing request loads and resource sharing on edge servers and Internet of Things devices.
- Demonstrate adaptive edge deep learning in real applications, especially elder care, with the Massachusetts AI and Technology Center for Connected Care in Aging and Alzheimer's Disease.
- Design application-level techniques that let multiple tasks on embedded systems share intermediate results, reducing memory use without hurting performance (MemDrive project).
- Develop system-level and cross-stack memory management that reduces duplicated memory across processes and uses predictable usage patterns, tested on robotic testbeds.
More about the research
These devices have little memory, computing power and energy. Both of this lab's projects ask how to build software systems that get the most out of limited hardware, so that larger and more capable AI models can run on small devices.
Why it matters
These systems could let affordable edge devices, robots and Internet of Things products run more powerful AI models than their hardware now allows. That could support new applications, including technology for elder care and people with Alzheimer's disease. Better use of memory in embedded systems could also make low-cost robotics and smart devices more capable and easier to obtain.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Lena Mashayekhy
University of Delaware, DE·Computer Science·Early-career
Develop a data-driven method to find mobility patterns of Internet of Things devices that lead to quality-of-service violations in multiaccess edge computing systems.
Shaahin Angizi
New Jersey Institute of Technology, NJ·Engineering·Early-career
Design reconfigurable circuits and architectures for a multi-modal in-sensor accelerator that moves AI computation onto sensor nodes as efficiently as possible.
Yingyan Lin
Georgia Institute of Technology, GA·Computer Science·Early-career
Create algorithms that summarize streaming data on each edge device to cut memory use and latency while preserving privacy, as the basis for decentralized incremental learning.
The school
University of Massachusetts: 28 labs on PI Pages, $282M a year on research, 337 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.