Shuai Zhang
New Jersey Institute of Technology, NJ·Engineering
Networks of connected sensors and devices, known as the Internet of Things (IoT), increasingly use artificial intelligence (AI) in settings such as smart health, power grids, robotics and multimedia systems. These AI-enabled IoT (AIoT) systems produce huge streams of data that change across both space and time.
- Fields
Engineering
Computer Science
- Topics
graph signal processing
graph neural networks
graph learning
- Grants
- 1 known grant, through Feb 2028
What they work on
- Develop topology sampling and graph neural network pruning methods for single-layer graph models, trimming network structure and model size so spatial-temporal AIoT data can be analyzed more efficiently.
- Investigate multilayer graph models, which stack several linked networks, to improve the processing of data that varies across space and time in AIoT systems.
- Analyze how signals propagate through AIoT networks and how graph structures evolve over time, using graph signal processing tools.
- Establish parameter-efficient transfer learning methods that adapt trained graph models to new spatial-temporal signal processing tasks while tuning only a small number of parameters.
More about the research
Examples include readings from many sensors in different locations recorded every moment. Analyzing this data efficiently, and in ways people can understand, affects how well and how cheaply these systems work.
Why it matters
Faster and more interpretable analysis of sensor data could improve AIoT applications such as smart health monitoring, power grid management, robotics and multimedia processing. It could also lower the computing cost of these systems. Engineers and operators who need to trust and understand AI decisions made from distributed sensor networks could benefit.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Georgios Giannakis
University of Minnesota, MN·Engineering
Model the resonant-beam optical wireless channel, an infrared beam formed in an open-air optical cavity, using nonlinear time-varying Volterra series and kernel-based learning for both downlink and uplink.
Jurij Leskovec
Stanford University, CA·Computer Science
Develop multi-level graph neural networks that handle graphs whose nodes are themselves complex structures, such as molecular graphs, for scientific foundation models.
My Thai
University of Florida, FL·Computer Science
Design communication topologies and schedules for Decentralized Federated Learning, where devices train a shared model by exchanging updates peer-to-peer, using graph theory and network measurements to improve convergence and robustness.
The school
New Jersey Institute of Technology: 12 labs on PI Pages, $172M a year on research, 74 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.