1. Meta-neural-network for real-time and passive deep-learning-based object recognition
- Author
-
Jingkai Weng, Chengbo Hu, Jing Yang, Bin Liang, Jianchun Cheng, Yujiang Ding, and Xue-Feng Zhu
- Subjects
FOS: Computer and information sciences ,Computer Science - Machine Learning ,Computer science ,Science ,Computer Science::Neural and Evolutionary Computation ,FOS: Physical sciences ,General Physics and Astronomy ,Applied Physics (physics.app-ph) ,02 engineering and technology ,01 natural sciences ,Article ,General Biochemistry, Genetics and Molecular Biology ,Machine Learning (cs.LG) ,0103 physical sciences ,Computer vision ,Neural and Evolutionary Computing (cs.NE) ,010306 general physics ,Multidisciplinary ,Artificial neural network ,Quantitative Biology::Neurons and Cognition ,business.industry ,Deep learning ,Computational science ,Cognitive neuroscience of visual object recognition ,Computer Science - Neural and Evolutionary Computing ,Metamaterial ,Physics - Applied Physics ,Acoustics ,General Chemistry ,021001 nanoscience & nanotechnology ,Object (computer science) ,Wave field ,Artificial intelligence ,Digit recognition ,0210 nano-technology ,business - Abstract
Analyzing scattered wave to recognize object is of fundamental significance in wave physics. Recently-emerged deep learning technique achieved great success in interpreting wave field such as in ultrasound non-destructive testing and disease diagnosis, but conventionally need time-consuming computer postprocessing or bulky-sized diffractive elements. Here we theoretically propose and experimentally demonstrate a purely-passive and small-footprint meta-neural-network for real-time recognizing complicated objects by analyzing acoustic scattering. We prove meta-neural-network mimics a standard neural network despite its compactness, thanks to unique capability of its metamaterial unit-cells (dubbed meta-neurons) to produce deep-subwavelength phase shift as training parameters. The resulting device exhibits the “intelligence” to perform desired tasks with potential to overcome the current limitations, showcased by two distinctive examples of handwritten digit recognition and discerning misaligned orbital-angular-momentum vortices. Our mechanism opens the route to new metamaterial-based deep-learning paradigms and enable conceptual devices automatically analyzing signals, with far-reaching implications for acoustics and related fields., The authors present a passive meta-neural-network for real-time recognition of objects by analysis of acoustic scattering. It consists of unit cells termed meta-neurons, mimicking an analogous neural network for classical waves, and is shown to recognise handwritten digits and misaligned orbital-angular-momentum vortices.
- Published
- 2020