QuATON: Quantization Aware Training of Optical Neurons

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Kariyawasam, Hasindu, Hettiarachchi, Ramith, Yang, Quansan, Matlock, Alex, Nambara, Takahiro, Kusaka, Hiroyuki, Kunai, Yuichiro, So, Peter T C, Boyden, Edward S, Wadduwage, Dushan
Formato: Preprint
Publicado: 2023
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913275437907968
author Kariyawasam, Hasindu
Hettiarachchi, Ramith
Yang, Quansan
Matlock, Alex
Nambara, Takahiro
Kusaka, Hiroyuki
Kunai, Yuichiro
So, Peter T C
Boyden, Edward S
Wadduwage, Dushan
author_facet Kariyawasam, Hasindu
Hettiarachchi, Ramith
Yang, Quansan
Matlock, Alex
Nambara, Takahiro
Kusaka, Hiroyuki
Kunai, Yuichiro
So, Peter T C
Boyden, Edward S
Wadduwage, Dushan
contents Optical processors, built with "optical neurons", can efficiently perform high-dimensional linear operations at the speed of light. Thus they are a promising avenue to accelerate large-scale linear computations. With the current advances in micro-fabrication, such optical processors can now be 3D fabricated, but with a limited precision. This limitation translates to quantization of learnable parameters in optical neurons, and should be handled during the design of the optical processor in order to avoid a model mismatch. Specifically, optical neurons should be trained or designed within the physical-constraints at a predefined quantized precision level. To address this critical issues we propose a physics-informed quantization-aware training framework. Our approach accounts for physical constraints during the training process, leading to robust designs. We demonstrate that our approach can design state of the art optical processors using diffractive networks for multiple physics based tasks despite quantized learnable parameters. We thus lay the foundation upon which improved optical processors may be 3D fabricated in the future.
format Preprint
id arxiv_https___arxiv_org_abs_2310_03049
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle QuATON: Quantization Aware Training of Optical Neurons
Kariyawasam, Hasindu
Hettiarachchi, Ramith
Yang, Quansan
Matlock, Alex
Nambara, Takahiro
Kusaka, Hiroyuki
Kunai, Yuichiro
So, Peter T C
Boyden, Edward S
Wadduwage, Dushan
Machine Learning
Image and Video Processing
Optics
Optical processors, built with "optical neurons", can efficiently perform high-dimensional linear operations at the speed of light. Thus they are a promising avenue to accelerate large-scale linear computations. With the current advances in micro-fabrication, such optical processors can now be 3D fabricated, but with a limited precision. This limitation translates to quantization of learnable parameters in optical neurons, and should be handled during the design of the optical processor in order to avoid a model mismatch. Specifically, optical neurons should be trained or designed within the physical-constraints at a predefined quantized precision level. To address this critical issues we propose a physics-informed quantization-aware training framework. Our approach accounts for physical constraints during the training process, leading to robust designs. We demonstrate that our approach can design state of the art optical processors using diffractive networks for multiple physics based tasks despite quantized learnable parameters. We thus lay the foundation upon which improved optical processors may be 3D fabricated in the future.
title QuATON: Quantization Aware Training of Optical Neurons
topic Machine Learning
Image and Video Processing
Optics
url https://arxiv.org/abs/2310.03049