🤖 AI Summary
This work addresses the challenges of high-speed, real-time multimedia signal processing by proposing a hierarchical photonic reservoir computing system that integrates binary optical modulation via a digital micromirror device (DMD), a random scattering medium, high-speed CMOS photodetection, and a time-multiplexed deep architecture to efficiently extract spatiotemporal features. By co-optimizing intra- and inter-layer physical hyperparameters, the system achieves a balance between memory retention and dynamic responsiveness, enabling the first scalable implementation of a hierarchical photonic reservoir. The proposed architecture demonstrates state-of-the-art performance across video, image, and speech recognition tasks, achieving processing rates exceeding the gigabit-per-second (Gbps) scale.
📝 Abstract
We present a deep photonic neural network architecture based on ultrafast binary optical modulation from a digital micro-mirror device (DMD), optical scattering in random medium, high-speed photodetection with a CMOS sensor, and time-multiplexed deep layer structure. Operating at Gigabit-per-second (Gb/s) processing rates, our system based on the reservoir computing (RC) framework achieves state-of-the-art performance across various multimedia tasks, including video, image and speech recognition. We show that the careful optimization of key physical intra- and inter-layer hyper-parameters can significantly enhance the deep photonic RC system ability to extract relevant temporal and spatial features via balancing memory retention and dynamical response of individual layers. This approach paves the way for highly scalable hierarchical photonic reservoir computing systems for high-throughput real-time multimedia signal processing.