{"title":"利用全光学衍射处理器的多路复用元表面进行多任务学习","authors":"Sahar Behroozinia, Qing Gu","doi":"arxiv-2409.08423","DOIUrl":null,"url":null,"abstract":"Diffractive Neural Networks (DNNs) leverage the power of light to enhance\ncomputational performance in machine learning, offering a pathway to\nhigh-speed, low-energy, and large-scale neural information processing. However,\nmost existing DNN architectures are optimized for single tasks and thus lack\nthe flexibility required for the simultaneous execution of multiple tasks\nwithin a unified artificial intelligence platform. In this work, we utilize the\npolarization and wavelength degrees of freedom of light to achieve optical\nmulti-task identification using the MNIST, FMNIST, and KMNIST datasets.\nEmploying bilayer cascaded metasurfaces, we construct dual-channel DNNs capable\nof simultaneously classifying two tasks, using polarization and wavelength\nmultiplexing schemes through a meta-atom library. Numerical evaluations\ndemonstrate performance accuracies comparable to those of individually trained\nsingle-channel, single-task DNNs. Extending this approach to three-task\nparallel recognition reveals an expected performance decline yet maintains\nsatisfactory classification accuracies of greater than 80% for all tasks. We\nfurther introduce a novel end-to-end joint optimization framework to redesign\nthe three-task classifier, demonstrating substantial improvements over the\nmeta-atom library design and offering the potential for future multi-channel\nDNN designs. Our study could pave the way for the development of ultrathin,\nhigh-speed, and high-throughput optical neural computing systems.","PeriodicalId":501214,"journal":{"name":"arXiv - PHYS - Optics","volume":"27 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Leveraging Multiplexed Metasurfaces for Multi-Task Learning with All-Optical Diffractive Processors\",\"authors\":\"Sahar Behroozinia, Qing Gu\",\"doi\":\"arxiv-2409.08423\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Diffractive Neural Networks (DNNs) leverage the power of light to enhance\\ncomputational performance in machine learning, offering a pathway to\\nhigh-speed, low-energy, and large-scale neural information processing. However,\\nmost existing DNN architectures are optimized for single tasks and thus lack\\nthe flexibility required for the simultaneous execution of multiple tasks\\nwithin a unified artificial intelligence platform. In this work, we utilize the\\npolarization and wavelength degrees of freedom of light to achieve optical\\nmulti-task identification using the MNIST, FMNIST, and KMNIST datasets.\\nEmploying bilayer cascaded metasurfaces, we construct dual-channel DNNs capable\\nof simultaneously classifying two tasks, using polarization and wavelength\\nmultiplexing schemes through a meta-atom library. Numerical evaluations\\ndemonstrate performance accuracies comparable to those of individually trained\\nsingle-channel, single-task DNNs. Extending this approach to three-task\\nparallel recognition reveals an expected performance decline yet maintains\\nsatisfactory classification accuracies of greater than 80% for all tasks. We\\nfurther introduce a novel end-to-end joint optimization framework to redesign\\nthe three-task classifier, demonstrating substantial improvements over the\\nmeta-atom library design and offering the potential for future multi-channel\\nDNN designs. Our study could pave the way for the development of ultrathin,\\nhigh-speed, and high-throughput optical neural computing systems.\",\"PeriodicalId\":501214,\"journal\":{\"name\":\"arXiv - PHYS - Optics\",\"volume\":\"27 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-12\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Optics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.08423\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Optics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.08423","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Leveraging Multiplexed Metasurfaces for Multi-Task Learning with All-Optical Diffractive Processors
Diffractive Neural Networks (DNNs) leverage the power of light to enhance
computational performance in machine learning, offering a pathway to
high-speed, low-energy, and large-scale neural information processing. However,
most existing DNN architectures are optimized for single tasks and thus lack
the flexibility required for the simultaneous execution of multiple tasks
within a unified artificial intelligence platform. In this work, we utilize the
polarization and wavelength degrees of freedom of light to achieve optical
multi-task identification using the MNIST, FMNIST, and KMNIST datasets.
Employing bilayer cascaded metasurfaces, we construct dual-channel DNNs capable
of simultaneously classifying two tasks, using polarization and wavelength
multiplexing schemes through a meta-atom library. Numerical evaluations
demonstrate performance accuracies comparable to those of individually trained
single-channel, single-task DNNs. Extending this approach to three-task
parallel recognition reveals an expected performance decline yet maintains
satisfactory classification accuracies of greater than 80% for all tasks. We
further introduce a novel end-to-end joint optimization framework to redesign
the three-task classifier, demonstrating substantial improvements over the
meta-atom library design and offering the potential for future multi-channel
DNN designs. Our study could pave the way for the development of ultrathin,
high-speed, and high-throughput optical neural computing systems.