Programmable phase-change metasurfaces on waveguides for multimode photonic convolutional neural network
Программируемые фазово-переводные метаповерхности на волноводах для многомодовой фотонной сверточной нейронной сети
2021-01-04
SCID: 54.1/st949rqz
Discuss with AI
matrix-vector multiplication (MVM)multimode photonic convolutional neural networkon-waveguide metasurfacesphase-change material Ge2Sb2Te5programmable phase-change metasurfaces
Figures from the paper
Abstract (AI)
Abstract Neuromorphic photonics has recently emerged as a promising hardware accelerator, with significant potential speed and energy advantages over digital electronics for machine learning algorithms, such as neural networks of various types. Integrated photonic networks are particularly powerful in performing analog computing of matrix-vector multiplication (MVM) as they afford unparalleled speed and bandwidth density for data transmission. Incorporating nonvolatile phase-change materials in integrated photonic devices enables indispensable programming and in-memory computing capabilities for on-chip optical computing. Here, we demonstrate a multimode photonic computing core consisting of an array of programable mode converters based on on-waveguide metasurfaces made of phase-change materials. The programmable converters utilize the refractive index change of the phase-change material Ge 2 Sb 2 Te 5 during phase transition to control the waveguide spatial modes with a very high precision of up to 64 levels in modal contrast. This contrast is used to represent the matrix elements, with 6-bit resolution and both positive and negative values, to perform MVM computation in neural network algorithms. We demonstrate a prototypical optical convolutional neural network that can perform image processing and recognition tasks with high accuracy. With a broad operation bandwidth and a compact device footprint, the demonstrated multimode photonic core is promising toward large-scale photonic neural networks with ultrahigh computation throughputs.
Key Findings
1
A multimode photonic computing core is demonstrated using an array of programmable on-waveguide metasurfaces made of phase-change material Ge2Sb2Te5.
2
Modal contrast encodes matrix elements with 6-bit resolution, supporting both positive and negative values for matrix-vector multiplication (MVM).
3
Programmable mode converters exploit refractive index changes during phase transitions to control waveguide spatial modes with up to 64 levels in modal contrast.
4
The device enables prototypical optical convolutional neural network operation for image processing and recognition with high accuracy.
5
The multimode photonic core offers broad operation bandwidth and a compact footprint, supporting potential large-scale photonic neural networks with ultrahigh computation throughput.
Research Object
Multimode photonic computing core consisting of an array of programmable on-waveguide phase-change metasurface mode converters (Ge2Sb2Te5) integrated on waveguides
Research Subject
Use of refractive-index switching in phase-change metasurface mode converters to programmably control waveguide spatial modes for representing matrix elements (6-bit, positive/negative, up to 64 modal contrast levels) and performing multimode photonic matrix–vector multiplication for convolutional neural network inference
Publication Details
Publication Date
2021-01-04
Journal
Publisher
ISSN
Cited by
384
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai6
Broadband transparent optical phase change materials for high-performance nonvolatile photonics2019
Broadband nonvolatile photonic switching based on optical phase change materials: beyond the classical figure-of-merit2017
Low-Loss and Broadband Nonvolatile Phase-Change Directional Coupler Switches2019
Nonvolatile All‐Optical 1 × 2 Switch for Chipscale Photonic Networks2016
Miniature Multilevel Optical Memristive Switch Using Phase Change Material2019
Low-Loss Integrated Photonic Switch Using Subwavelength Patterned Phase Change Material2018
Cited by15
Optical meta-waveguides for integrated photonics and beyond2021
Monolithic back-end-of-line integration of phase change materials into foundry-manufactured silicon photonics2024
Opportunities and Challenges for Large-Scale Phase-Change Material Integrated Electro-Photonics2022
Non-Volatile Reconfigurable Silicon Photonics Based on Phase-Change Materials2021
Non‐Volatile Optical Switch Element Enabled by Low‐Loss Phase Change Material2023
Fabrication and integration of photonic devices for phase-change memory and neuromorphic computing2023
On-Chip Integrated Photonic Devices Based on Phase Change Materials2021
Nonvolatile Multilevel Switching of Silicon Photonic Devices with In 2 O 3 /GST Segmented Structures2023
Resonant multilevel optical switching with phase change material GST2022
Phase change material enabled 2 × 2 silicon nonvolatile optical switch2021
Advances in Photonic Devices Based on Optical Phase-Change Materials2021
Neuromorphic Photonics Based on Phase Change Materials2023
Fabrication of Smart Materials Using Laser Processing: Analysis and Prospects2023
Ultra-compact integrated photonic switch using miniature phase change material inserted in a slot waveguide2023
The changing face of phase change photonics2026