This presentation was made by Ade Clinton Sitepu – 110999406 for Final Project of Optical Communication course. If MEMS is the result of applying modern nanoscale CMOS processes to the mechanical world, then doing the same for the optical domain gives us Silicon photonics. According to the Moore’s Law, our computer capabilities also was doubling every 2-year. The problems on today’s Modern Era is the usage computational power to train Neural Networks was doubling every 3.4 months. The current situation cannot just satisfy this compute power that’s required for these type of AI based systems. So, there has been a need to go from software-based Neural Network on standard digital computers, to more specialized computers which we called Neuromorphic Hardware. It’s like enabling Neural Networks into Silicon Photonics chip. We have numbers of application on neuromorphic hardware and we will talk one of them in fiber optic communications and I will try explain my best to explain this.
