Research themes spanning space perception, explainable AI, neural representations, and scientific machine learning.
About the Lab
The NEural TransmissionS (NETS) Lab builds machine learning systems for scientific and engineering problems where standard methods break down. Our work spans space perception and autonomy, edge deployment, explainable AI and neural representations, and applied machine learning, with a distinctive grounding in applied mathematics.
A central application area is space-based perception: inspection, tracking, navigation, and 3D reconstruction of spacecraft in challenging orbital environments. These are safety-critical settings, where models must generalize reliably, be understood, and run onboard under strict power, memory, latency, and computational constraints.
Research Perspective
NETS approaches machine learning as both a mathematical and systems problem. We are interested not only in performance, but in what models are actually learning, how their internal structure evolves, and how that behavior can be shaped through theory, regularization, and objective design.
Collaborative Model
The lab works alongside collaborators from government and defense laboratories, commercial partners, and academic groups across disciplines. Projects often connect theory, modeling, simulation, embedded implementation, and deployment, with collaborations spanning areas such as aerospace, biology, medicine, weather, and engineering.
Training Environment
The lab includes contributors at multiple stages, from high school students and undergraduates through master's and PhD researchers, often working on connected problems across different time scales. The goal is to train researchers who can move comfortably between foundations, experiments, edge implementation, and real deployment constraints.