Inference and Gradient Measurement for Backpropagation in Photonic Neural Networks

Pai, Sunil; Hughes, Tyler W.; Park, Tae-Won; Bartlett, Ben; Williamson, Ian A. D.; Minkov, Momchil; Milanizadeh, Maziyar; Abebe, Nathnael; Morichetti, Francesco; Melloni, Andrea; Solgaard, Olav; Fan, Shanhui; Miller, D. A. B.

doi:10.1364/cleo_si.2022.sth5g.2

Cited by 8 publications

(13 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such process involving hardware physics and imperfection in the training loop can be broadly applicable to various physical systems, such as electronic, optical, and mechanical systems. [ 48–55 ] Furthermore, we show that the developed O‐GEMM hardware accelerator can be further employed in the RL algorithms to accelerate the chip design of another optical accelerator.…”

Section: Resultsmentioning

confidence: 99%

Device‐System End‐to‐End Design of Photonic Neuromorphic Processor Using Reinforcement Learning

Tang

Zamani

Chen

et al. 2022

Laser & Photonics Reviews

View full text Add to dashboard Cite

The incorporation of high-performance optoelectronic devices into photonic neuromorphic processors can substantially accelerate computationally intensive matrix multiplication operations in machine learning (ML) algorithms. However, the conventional designs of individual devices and system are largely disconnected, and the system optimization is limited to the manual exploration of a small design space. Here, a device-system end-to-end design methodology is reported to optimize a free-space optical general matrix multiplication (GEMM) hardware accelerator by engineering a spatially reconfigurable array made from chalcogenide phase change materials. With a highly parallelized integrated hardware emulator with experimental information, the design of unit device to directly optimize GEMM calculation accuracy is achieved by exploring a large parameter space through reinforcement learning algorithms, including deep Q-learning neural network, Bayesian optimization, and their cascaded approach. The algorithm-generated physical quantities show a clear correlation between system performance metrics and device specifications. Furthermore, physics-aware training approaches are employed to deploy optimized hardware to the tasks of image classification, materials discovery, and a closed-loop design of optical ML accelerators. The demonstrated framework offers insights into the end-to-end and co-design of optoelectronic devices and systems with reduced human supervision and domain knowledge barriers.

show abstract

Section: Resultsmentioning

confidence: 99%

Device‐System End‐to‐End Design of Photonic Neuromorphic Processor Using Reinforcement Learning

Tang

Zamani

Chen

et al. 2022

Laser & Photonics Reviews

View full text Add to dashboard Cite

show abstract

“…We have demonstrated PICs consisting of a full package of elementary photonic components, including waveguides, grating couplers, ring resonators, MZIs, programmable optical switch fabrics, reconfigurable photonic crossbar array, and tunable optical filters. Although our demonstrations have been conducted in a near-in situ fashion, in steps of writing, measuring, modifying, and checking, real-time reconfiguration and feedback-controlled adaptation (6,30,49,62) of the PICs are entirely feasible. Furthermore, the application scenario can be expanded by introducing a multilevel grayscale design (37,39,57,(63)(64)(65) instead of the current binary design or by selecting appropriate substrates tailored to specific applications.…”

Section: Discussionmentioning

confidence: 99%

“…We have demonstrated PICs consisting of a full package of elementary photonic components, including waveguides, grating couplers, ring resonators, MZIs, programmable optical switch fabrics, reconfigurable photonic crossbar array, and tunable optical filters. Although our demonstrations have been conducted in a near–in situ fashion, in steps of writing, measuring, modifying, and checking, real-time reconfiguration and feedback-controlled adaptation ( 6 , 30 , 49 , 62 ) of the PICs are entirely feasible.…”

Section: Discussionmentioning

confidence: 99%

Freeform direct-write and rewritable photonic integrated circuits in phase-change thin films

Wu,

Deng,

Huang

et al. 2024

Sci. Adv.

View full text Add to dashboard Cite

Photonic integrated circuits (PICs) with rapid prototyping and reprogramming capabilities promise revolutionary impacts on a plethora of photonic technologies. We report direct-write and rewritable photonic circuits on a low-loss phase-change material (PCM) thin film. Complete end-to-end PICs are directly laser-written in one step without additional fabrication processes, and any part of the circuit can be erased and rewritten, facilitating rapid design modification. We demonstrate the versatility of this technique for diverse applications, including an optical interconnect fabric for reconfigurable networking, a photonic crossbar array for optical computing, and a tunable optical filter for optical signal processing. By combining the programmability of the direct laser writing technique with PCM, our technique unlocks opportunities for programmable photonic networking, computing, and signal processing. Moreover, the rewritable photonic circuits enable rapid prototyping and testing in a convenient and cost-efficient manner, eliminate the need for nanofabrication facilities, and thus promote the proliferation of photonics research and education to a broader community.

show abstract

“…The energy consumption of PAGT method is 62 J, which is significantly lower than 280 J for the in situ method. [27] From the analysis of the scalability, our PAGT retains its advantages on energy and time consumption for most of the current PNNs. Besides, in some special network architectures that require the derivatives of hybrid photonic-digital models, [28,29] unifying the non-differentiable photonic part and the derivatives of the differential digital part into analytic derivatives will greatly accelerate the training speed and reduce training costs.…”

Section: Introductionmentioning

confidence: 99%

Physics‐Aware Analytic‐Gradient Training of Photonic Neural Networks

Zhan,

Zhang,

Lin

et al. 2024

Laser & Photonics Reviews

View full text Add to dashboard Cite

Photonic neural networks (PNNs) have emerged as promising alternatives to traditional electronic neural networks. However, the training of PNNs, especially the chip implementation of analytic gradient descent algorithms that are recognized as highly efficient in traditional practice, remains a major challenge because physical systems are not differentiable. Although training methods such as gradient‐free and numerical gradient methods are proposed, they suffer from excessive measurements and limited scalability. State‐of‐the‐art in situ training method is also cost‐challenged, requiring expensive in‐line monitors and frequent optical I/O switching. Here, a physics‐aware analytic‐gradient training (PAGT) method is proposed that calculates the analytic gradient in a divide‐and‐conquer strategy, overcoming the difficulty induced by chip non‐differentiability in the training of PNNs. Multiple training cases, especially a generative adversarial network, are implemented on‐chip, achieving a significant reduction in time consumption (from 31 h to 62 min) and a fourfold reduction in energy consumption, compared to the in situ method. The results provide low‐cost, practical, and accelerated solutions for training hybrid photonic‐digital electronic neural networks.

show abstract

Inference and Gradient Measurement for Backpropagation in Photonic Neural Networks

Cited by 8 publications

References 2 publications

Device‐System End‐to‐End Design of Photonic Neuromorphic Processor Using Reinforcement Learning

Device‐System End‐to‐End Design of Photonic Neuromorphic Processor Using Reinforcement Learning

Freeform direct-write and rewritable photonic integrated circuits in phase-change thin films

Physics‐Aware Analytic‐Gradient Training of Photonic Neural Networks

Contact Info

Product

Resources

About