.png){kind=icon}
Return
Efficient stochastic parallel gradient descent training for on-chip optical processor
.png){kind=icon}
OA
AI
DOI：10.29026/oea.2024.230182.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In recent years, space -division multiplexing (SDM) technology, which involves transmitting data information on multiple parallel channels for efficient capacity scaling, has been widely used in fiber and free -space optical communication tems. To enable flexible data management and cope with the mixing between different channels, the integrated reconfigurable optical processor is used for optical switching and mitigating the channel crosstalk. However, efficient online training becomes intricate and challenging, particularly when dealing with a significant number of channels. Here we use stochastic parallel gradient descent (SPGD) algorithm to configure the integrated optical processor, which has less computation than the traditional gradient descent (GD) algorithm. We design and fabricate a 6x6 on -chip optical processor silicon platform to implement optical switching and descrambling assisted by the online training with the SPDG algorithm. Moreover, we apply the on -chip processor configured by the SPGD algorithm to optical communications for optical switching and efficiently mitigating the channel crosstalk in SDM systems. In comparison with the traditional GD gorithm, it is found that the SPGD algorithm features better performance especially when the scale of matrix is large, which means it has the potential to optimize large-scale optical matrix computation acceleration chips.
Keywords:
optical communications
optical signal processing
channel descrambling
optical neural network chip
silicon photonics
Journal
IF:
22.4
Papers:
413
Citations:
3.6K
Organization
No organization information available
Cited Papers
A Compact Butterfly-Style Silicon Photonic-Electronic Neural Chip for Hardware-Efficient Deep Learning.png)
ACS PHOTONICSOAAI
ACS PHOTONICSIF6.7
OAAIShareSaveShareSaveShareSaveShareSave18 km low-crosstalk OAM plus WDM transmission with 224 individual channels enabled by a ring-core fiber with large high-order mode group separationOPTICS LETTERS
OPTICS LETTERSIF3.3
PREAI
ShareSaveExperimental demonstration of using divergence cost-function in SPGD algorithm for coherent beam combining with tip/tilt controlOPTICS EXPRESS
OAAI
OPTICS EXPRESSIF3.3
OAAIShareSaveUpstream ORFs Prevent MAVS Spontaneous Aggregation and Regulate Innate Immune HomeostasisiScienceOAAI
iScienceIF0
OAAIShareSave
View more