Pretty mind-blowing paper, 3D-printed representations of neural networks you can run inference on by shining light through... literal *light speed* inference times.— Robbie Barrat (@DrBeef_) August 1, 2018
They only have classification networks so far, but I'd love to see pix2pix or a GAN soon.https://t.co/mTqIzwQTEK pic.twitter.com/bTI0yXvT5F
Xing Lin, Yair Rivenson, Nezih T. Yardimci, Muhammed Veli, Yi Luo, Mona Jarrahi1, Aydogan Ozcan, All-optical machine learning using diffractive deep neural networks, Science 26 Jul 2018: eaat8084, DOI: 10.1126/science.aat8084
Abstract: Deep learning has been transforming our ability to execute advanced inference tasks using computers. We introduce a physical mechanism to perform machine learning by demonstrating an all-optical Diffractive Deep Neural Network (D2NN) architecture that can implement various functions following the deep learning-based design of passive diffractive layers that work collectively. We create 3D-printed D2NNs that implement classification of images of handwritten digits and fashion products as well as the function of an imaging lens at terahertz spectrum. Our all-optical deep learning framework can perform, at the speed of light, various complex functions that computer-based neural networks can implement, and will find applications in all-optical image analysis, feature detection and object classification, also enabling new camera designs and optical components that perform unique tasks using D2NNs.
This resonates with some work I read years ago:
Abu-Mostafa, Y. & Psaltis, D. (1987) Optical Neural Computers. Scientific American, 256, 3, pp. 88-95.
Psaltis, D. & Mok, F. (1995) Holographic Memories. Scientific American 273, 5, pp. 70-76.