All-optical machine learning using diffractive deep neural networks
By Xing Lin, Yair Rivenson, Nezih T. Yardimci, Muhammed Veli, Yilin Luo, Mona Jarrahi, Aydogan Ozcan. 2018
2,293 citations recorded by the source corpus.
All-optical deep learning Deep learning uses multilayered artificial neural networks to learn digitally from large datasets. It then performs advanced identification and classification tasks. To date, these multilayered neural networks have been implemented on a computer. Lin et al. demonstrate all-optical machine learning that uses passive optical components that can be patterned and fabricated with 3D-printing. Their hardware approach comprises stacked layers of diffractive optical elements analogous to an artificial neural network that can be trained to execute complex functions at the speed of light. Science, this issue p. 1004 All-optical deep learning can be implemented with 3D-printed passive optical components. Deep learning has been transforming our ability to execute advanced inference tasks using computers. Here 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 created 3D-printed D2NNs that implement classification of images of handwritten digits and fashion products, as well as the function of an imaging lens at a terahertz spectrum. Our all-optical deep learning framework can perform, at the speed of light, various complex functions that computer-based neural networks can execute; will find applications in all-optical image analysis, feature detection, and object classification; and will also enable new camera designs and optical components that perform distinctive tasks using D2NNs.
Subfields: Artificial Intelligence, Electrical and Electronic Engineering, ML Systems & Hardware, AI for Physical Sciences
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Peter Leonard McMahon
- Andy Zeng
- Yun Raymond Fu
- Sara Hooker
- Sergey Levine
- Vikas Sindhwani
- Marin Soljačić
- Peter R. Florence
In Electrical and Electronic Engineering
- Peter Leonard McMahon
- Marin Soljačić
- Dirk Englund
- Shanhui Fan
- Ryan Hamerly
- Alexander Sludds
- Liane Bernstein
- David A. B. Miller
In ML Systems & Hardware
In AI for Physical Sciences
Citations in the corpus
- Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language (2022, background): “…with fixed (pretrained) models zero-shot, this may also help foster adoption of new machine learning accelerators (e.g., fixed analog circuity (Reuther et al. 2020), optical diffraction (Lin et al. 2018)) for inference with substantially lower power consumption and more compact form factors.”
- The hardware lottery (2020, background): “…include biological hardware (Tan et al., 2007; Macía & Sole, 2014; Kriegman et al., 2020), analog hardware with in-memory computation (Ambro-gio et al., 2018), neuromorphic computing (Davies, 2019), optical computing (Lin et al., 2018), and quantum computing based approaches (Cross et al., 2019).”
- Deep Learning in Mobile and Wireless Networking: A Survey (2018, background): “Diffractive neural networks (D2NNs) that completely rely on light communication were recently introduced in [117],”
- Machine learning and applications in ultrafast photonics (2020, methodology): “on multiple layers of diffractive 369 surfaces where each point on a given layer acts as a node [96], or based on optical matrix 370 multiplication using a cascaded array of Mach–Zehnder interferometers integrated into a 371 silicon photonic circuit [97].”
- Stacked Intelligent Metasurfaces for Efficient Holographic MIMO Communications in 6G (2023, neutral): “Moreover, the transmission coefficient from the n-th metaatom on the k-th receive metasurface layer to the ñ-th metaatom on the (k − 1)-st receive metasurface layer is expressed by [46]”
- Inference in artificial intelligence with deep optics and photonics (2020, methodology): “With such layers, several optimized phase or amplitude patterns can be mounted with some distance between them to implement an all-optical classification algorithm 14 .”
- Photonic matrix multiplication lights up photonic accelerator and beyond (2022, methodology): “Afterwards, the MPLC technique was widely used in various fi elds, such as for all-optical machine learning 46 – 48 , the Laguerre-Gaussian or orbital angular momentum (OAM) mode sorter 49,50 , the photonic Ising machine 51,52 , time-reversed optical waves 53 , optical logic operations 54 , optical…”
- Deep Neural Network Inverse Design of Integrated Photonic Power Splitters (2019, methodology): “It has received some attention in optical community and there has been several recent work on reverse modeling for design of nano-structured optical components using DNN [19-25], as well as hardware implementation of an artificial neural network [26-30].”
- Deep physical neural networks trained with backpropagation (2021, methodology): “16, 17), it is well suited to the MNIST handwritten digit classification task, achieving nearly the same performance as a digital linear model 55 .”