Zobrazeno 1 - 5
of 5
pro vyhledávání: '"Darren T. Schachter"'
Autor:
Logan G. Wright, Tatsuhiro Onodera, Martin M. Stein, Tianyu Wang, Darren T. Schachter, Zoey Hu, Peter L. McMahon
Publikováno v:
AI and Optical Data Sciences IV.
We experimentally demonstrate multilayer neural networks using ultrafast nonlinear optics, to perform audio and image classification. The proposed framework for constructing and training neural networks is general and applicable to other complex non-
Autor:
Logan G. Wright, Tatsuhiro Onodera, Martin M. Stein, Tianyu Wang, Darren T. Schachter, Zoey Hu, Peter L. McMahon
Publikováno v:
Nature
Deep-learning models have become pervasive tools in science and engineering. However, their energy requirements now increasingly limit their scalability1. Deep-learning accelerators2–9 aim to perform deep learning energy-efficiently, usually target
Autor:
Logan G. Wright, Tatsuhiro Onodera, Martin M. Stein, Tianyu Wang, Darren T. Schachter, Zoey Hu, Peter L. McMahon
Publikováno v:
Conference on Lasers and Electro-Optics.
We have shown that nonlinear pulse propagation can be designed to directly, all-optically implement machine learning calculations. Building from proof-of-concept experiments, we analyze opportunities for ultrafast, ultra-efficient nonlinear optical n
Autor:
Logan G. Wright, Tatsuhiro Onodera, Martin M. Stein, Tianyu Wang, Darren T. Schachter, Zoey Hu, Peter L. McMahon
Publikováno v:
OSA Nonlinear Optics 2021.
We experimentally demonstrate an optical neural network whose fundamental building blocks are controllable ultrafast quadratic nonlinear pulse-propagation processes. The network is trained in situ by a new backpropagation algorithm that generalizes t
Autor:
Tianyu Wang, Darren T. Schachter, Zoey Hu, Martin M. Stein, Logan G. Wright, Peter L. McMahon, Tatsuhiro Onodera
Publikováno v:
Conference on Lasers and Electro-Optics.
We experimentally demonstrate deep nonlinear optical neural networks using a universal algorithm for backpropagating through arbitrary physical input-output transformations. Ultrafast second harmonic generation and other diverse processes are trained