Graduate Research Assistant

Research

I developed a hyperdimensional computing algorithm for probabilistic occupancy grid mapping that cut memory complexity by one order of magnitude and computational complexity by two orders of magnitude against existing approaches.

Funded work

I co-authored the technical approach and generated the preliminary results cited in support of two awarded US Army DEVCOM Automotive Research Center (ARC) proposals. Maryam Parsa was the principal investigator, and I was the named student researcher on each award. ARC is sponsored by the DEVCOM Ground Vehicle Systems Center (GVSC) for ground-vehicle modeling and simulation.

• Multi-Phase Vector Symbolic Architectures for Distributed and Collective Intelligence in Multi-Agent Autonomous Systems — $384,000 (GMU budget), January 2023 – May 2026.
• NEMOSYS: Neural Memory Organization System — $134,750 (GMU budget), May 2025 – December 2026; initially awarded in 2025 and extended through December 2026.

Tooling and collaboration

• Built the Event-based Sensing Algorithm Development Environment with the Leidos Innovations Center.
• First external contributor to Intel’s Lava-Optimization library.

Selected projects and papers

• Spatial representation and memory: VSA-OGM project and journal article; SRMU on streaming memory updates; HyperSpace and qFHRR preprints on hyperdimensional representations.
• Neuromorphic optimization: Neuromorphic Bayesian Optimization in Lava and Parallelized Multi-Agent Bayesian Optimization in Lava. Both publication entries link to the Lava-Optimization code.
• Graph learning: Transductive Spiking Graph Neural Networks for Loihi and VS-Graph.
• Swarm simulation and control: Zespol, Spiking Neural Networks as a Controller for Emergent Swarm Agents, and A Milling Swarm of Ground Robots using Spiking Neural Networks.
• Event-based vision: Retina-inspired Object Motion Segmentation for Event-Cameras and Hardware-Algorithm Re-engineering of Retinal Circuit for Intelligent Object Motion Segmentation.