The award comes from the NSF’s Cyber-Physical Systems program and will support Sivakumar and Fekri’s project titled CPS: Small: Multi-Human Assisted Learning for Multi-Agent Systems using Intrinsically Generated Event-Related EEG Potentials.

Machine learning solutions are particularly useful for monitoring, instrumenting, and optimizing complex cyber-physical systems (CPS). Supervised or controlled by computer-based algorithms, CPS’s include things like autonomous cars, smart grids, and automatic pilot avionics. However, these systems still need human help recognizing natural speech patterns and complex scenes or images.

Sivakumar and Fekri’s goal is to allow humans to assist with machine learning algorithms in this environment. Their approach uses a human-in-the-loop model. Using electroencephalogram (EEG), the team hopes to use human brain wave activity to help speed up the learning process for the algorithms that power CPS’s.

The results of the research will be evaluated using both game proxy-based analysis and experimental analysis within the context of real-world CPS.

The team’s award runs for three years beginning in January 2019. The team hopes to explore commercialization of the outcomes of the research in collaboration with Georgia Tech’s CREATE-X and Venture Lab programs.