Intellectual Merit: Real-time high capacity signal processing is very challenging to today's signal processing technologies; however, this kind of powerful processing paradigm is critical to fulfilling the increasing demand of large bandwidth signals in various applications that can no longer be supported by electronics, including biomedical, military, and consumer applications. Therefore, there is a critical need to investigate a novel approach to tackle these challenges. This BRIGE project borrows the principle of the Leaky Integrate-and-Fire Neuron and of Spike Timing Dependent Plasticity-based learning algorithm from physiological circuitry, and implements it with photonics. The proposed approach is an interdisciplinary integrated device and systems design study involving basic research from the standpoint of neuromorphic processing. The project aims to exploit physical processes in photonic devices, and to solve the problem of fundamental real-time learning and adaptive control in a dynamic system. This fundamentally novel approach is a neural-inspired way to solve the problem by eliminating the computing bottleneck and providing the high fidelities and large bandwidth processing ability needed for high-capacity and high-bandwidth signal processing.
Broader Impacts: The educational impact of the proposed work comes from its multi-disciplinary foundation, broadening students' views and encouraging them to think creatively. The multi-disciplinary nature of the project offers a wide array of exciting opportunities for developing new educational methods that can teach a new generation of students to wield neuroscience, photonics, and signal processing methods to attack problems. Toward this end, the investigating team plans to focus its educational efforts on redesigning fiber optics classes, encouraging high school and undergraduate students to participate in research, and engaging industrial partners in continuing educational activities. In particular, the team would like to increase opportunities for both minority and female students to get involved in research, and to provide mentoring throughout their early career development. The long-term societal benefits of the project originate from the high-capacity, real-time, and unique processing algorithm provided by photonic implementation of biological neurons. Self-learning and adaptive systems provide accurate and real-time adaptive capability in sensing systems, which can significantly reduce avoidable and unnecessary death and property damage due to natural and manmade disasters. They can also improve national security by enabling more dynamic and reliable high-capacity communication channels for the military, as well as forming a fast and accurate processing paradigm for identifying intentional terrorist actions. The unique properties of photonic neuron form the foundation for a low false alarm rate health condition alert system.
