Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks
The slowdown of Moore’s law and the existence of the “von Neumann bottleneck” has led to electronic-based computing systems under von Neumann’s architecture being unable to meet the fast-growing demand for artificial intelligence computing. However, all-optical diffractive neural networks provide a...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-12-01
|
| Series: | Micromachines |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-666X/15/1/50 |
| _version_ | 1797342967660281856 |
|---|---|
| author | Huayi Sheng Muhammad Shemyal Nisar |
| author_facet | Huayi Sheng Muhammad Shemyal Nisar |
| author_sort | Huayi Sheng |
| collection | DOAJ |
| description | The slowdown of Moore’s law and the existence of the “von Neumann bottleneck” has led to electronic-based computing systems under von Neumann’s architecture being unable to meet the fast-growing demand for artificial intelligence computing. However, all-optical diffractive neural networks provide a possible solution to this challenge. They can outperform conventional silicon-based electronic neural networks due to the significantly higher speed of the propagation of optical signals (≈<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mn>10</mn><mn>8</mn></msup></semantics></math></inline-formula> m.s<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mrow><mo>−</mo><mn>1</mn></mrow></msup></semantics></math></inline-formula>) compared to electrical signals (≈<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mn>10</mn><mn>5</mn></msup></semantics></math></inline-formula> m.s<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mrow><mo>−</mo><mn>1</mn></mrow></msup></semantics></math></inline-formula>), their parallelism in nature, and their low power consumption. The integrated diffractive deep neural network (ID<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mn>2</mn></msup></semantics></math></inline-formula>NN) uses an on-chip fully passive photonic approach to achieve the functionality of neural networks (matrix–vector operations) and can be fabricated via the CMOS process, which is technologically more amenable to implementing an artificial intelligence processor. In this paper, we present a detailed design framework for the integrated diffractive deep neural network and corresponding silicon-on-insulator integration implementation through Python-based simulations. The performance of our proposed ID<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mn>2</mn></msup></semantics></math></inline-formula>NN was evaluated by solving image classification problems using the MNIST dataset. |
| first_indexed | 2024-03-08T10:40:52Z |
| format | Article |
| id | doaj.art-edfd33affa3344a69089d0955ca50971 |
| institution | Directory Open Access Journal |
| issn | 2072-666X |
| language | English |
| last_indexed | 2024-03-08T10:40:52Z |
| publishDate | 2023-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Micromachines |
| spelling | doaj.art-edfd33affa3344a69089d0955ca509712024-01-26T17:43:27ZengMDPI AGMicromachines2072-666X2023-12-011515010.3390/mi15010050Simulating an Integrated Photonic Image Classifier for Diffractive Neural NetworksHuayi Sheng0Muhammad Shemyal Nisar1Sino-British College, University of Shanghai for Science and Technology, Shanghai 200093, ChinaSino-British College, University of Shanghai for Science and Technology, Shanghai 200093, ChinaThe slowdown of Moore’s law and the existence of the “von Neumann bottleneck” has led to electronic-based computing systems under von Neumann’s architecture being unable to meet the fast-growing demand for artificial intelligence computing. However, all-optical diffractive neural networks provide a possible solution to this challenge. They can outperform conventional silicon-based electronic neural networks due to the significantly higher speed of the propagation of optical signals (≈<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mn>10</mn><mn>8</mn></msup></semantics></math></inline-formula> m.s<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mrow><mo>−</mo><mn>1</mn></mrow></msup></semantics></math></inline-formula>) compared to electrical signals (≈<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mn>10</mn><mn>5</mn></msup></semantics></math></inline-formula> m.s<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mrow><mo>−</mo><mn>1</mn></mrow></msup></semantics></math></inline-formula>), their parallelism in nature, and their low power consumption. The integrated diffractive deep neural network (ID<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mn>2</mn></msup></semantics></math></inline-formula>NN) uses an on-chip fully passive photonic approach to achieve the functionality of neural networks (matrix–vector operations) and can be fabricated via the CMOS process, which is technologically more amenable to implementing an artificial intelligence processor. In this paper, we present a detailed design framework for the integrated diffractive deep neural network and corresponding silicon-on-insulator integration implementation through Python-based simulations. The performance of our proposed ID<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mrow></mrow><mn>2</mn></msup></semantics></math></inline-formula>NN was evaluated by solving image classification problems using the MNIST dataset.https://www.mdpi.com/2072-666X/15/1/50integrated photonicscomputing metasurfacesdiffractive neural networksphotonic image classifier |
| spellingShingle | Huayi Sheng Muhammad Shemyal Nisar Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks Micromachines integrated photonics computing metasurfaces diffractive neural networks photonic image classifier |
| title | Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks |
| title_full | Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks |
| title_fullStr | Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks |
| title_full_unstemmed | Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks |
| title_short | Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks |
| title_sort | simulating an integrated photonic image classifier for diffractive neural networks |
| topic | integrated photonics computing metasurfaces diffractive neural networks photonic image classifier |
| url | https://www.mdpi.com/2072-666X/15/1/50 |
| work_keys_str_mv | AT huayisheng simulatinganintegratedphotonicimageclassifierfordiffractiveneuralnetworks AT muhammadshemyalnisar simulatinganintegratedphotonicimageclassifierfordiffractiveneuralnetworks |