BRIGE: LEON: Learning in Optical Neuron

ofk, mable, university of georgie: Learning Optical Neuron （LEON）摘要智能优点：实时高容量信号处理对当今的信号处理技术提出了很大的挑战；然而，这种强大的处理范例对于满足各种应用中对大带宽信号日益增长的需求至关重要，这些应用不再由电子产品支持，包括生物医学，军事和消费应用。因此，迫切需要研究一种新的方法来应对这些挑战。该bridge项目借鉴了生理电路中的漏积分-放电神经元和基于峰值时序依赖的可塑性学习算法的原理，并利用光子学实现了它。所提出的方法是一种跨学科的集成设备和系统设计研究，涉及从神经形态处理的角度进行基础研究。该项目旨在利用光子器件中的物理过程，解决动态系统中基本的实时学习和自适应控制问题。这种新颖的方法是一种神经启发的方法，通过消除计算瓶颈，提供高容量和高带宽信号处理所需的高保真度和大带宽处理能力来解决问题。更广泛的影响：建议的工作的教育影响来自于它的多学科基础，拓宽学生的视野，鼓励他们创造性地思考。该项目的多学科性质为开发新的教育方法提供了一系列令人兴奋的机会，这些方法可以教会新一代学生运用神经科学、光子学和信号处理方法来解决问题。为此，调查小组计划将教育工作重点放在重新设计光纤课程，鼓励高中生和大学生参与研究，并让工业合作伙伴参与继续教育活动上。特别是，该团队希望增加少数民族和女性学生参与研究的机会，并在他们的早期职业发展中提供指导。该项目的长期社会效益源于生物神经元光子实现提供的高容量，实时和独特的处理算法。自学习和自适应系统为传感系统提供了准确、实时的自适应能力，可以显著减少自然灾害和人为灾害造成的可避免和不必要的死亡和财产损失。它们还可以通过为军队提供更动态、更可靠的高容量通信渠道，以及形成快速、准确的处理范式来识别故意的恐怖主义行动，从而改善国家安全。光子神经元的独特特性为低虚警率的健康状况报警系统奠定了基础。

ECCS-1342177Fok, MableUniversity of GeorgiaBRIGE: Learning Optical Neuron (LEON)ABSTRACTIntellectual 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.