Deep neural networks (DNNs) have shown their superiority in a variety of complicated machine learning tasks. However, large-scale DNNs are computation- and memory-intensive, and significant efforts have been made to improve the efficiency of DNNs through the use of better hardware accelerators as well as software training algorithms. The optical neural network (ONN) is a promising candidate as a next-generation neurocomputing platform due to its high parallelism, low latency, and low energy consumption. Here, we devise a hardware-efficient optical neural network architecture named optical subspace neural network (OSNN), which targets lower optical component usage, area cost, and energy consumption of previous ONN architectures with comparable task performance. Additionally, a hardware-aware training framework is provided to minimize the required control precision, lessen the chip area, and boost the noise robustness.
This Conference Presentation, “Wavelength-division-multiplexing-based electronic-photonic integrated circuits for high-performance data processing and transportation,” was recorded for the Photonics West 2021 Digital Forum.
This Conference Presentation, “Scalable fast-Fourier-transform-based (FFT-based) integrated optical neural network for compact and energy-efficient deep learning,” was recorded for the Photonics West 2021 Digital Forum.
This Conference Presentation, "Wavelength-division-multiplexing-based electronic-photonic network for high-speed computing" was recorded at Photonics West 2020 held in San Francisco, California, United States.
As a tremendous amount of data is being created exponentially day by day, integrated optical computing starts to attract lots of attention recently due to the bottleneck in the continuation of Moore’s law. With the rapid development of micro/nano-scale optical devices, integrated photonics has shown its potential to satisfy the demand of computation with an ultracompact size, ultrafast speed, and ultralow power consumption. As one of the paradigms in optical computing, the electro-optic logic that combines the merits of photonics and electronics has made considerable progress in various fundamental logic gates. It therefore becomes very critical to develop an automated design method to synthesize these logic devices for large-scale optical computing circuits. In this paper, we propose a new automated logic synthesis algorithm based on And-Inverter Graphs (AIGs) for electro-optic computing. A comprehensive component library of electro-optic logic is summarized with several new proposed logic gates. As an example, a large-scale ripple-carry full adder which serves as the core part of the arithmetic logical unit (ALU) is presented. In the design, all the electrical signals could be applied simultaneously at every clock cycle and then the light could process the signals through every bit at the speed of light without any delay accumulated. High-speed experiment demonstrations are carried out, which show its potential in future high-speed and low-power-consumption optical computing.
Due to ever increasing demand for information bandwidth, and with electronics approaching their performance limit, there has been a renewed interest in using optical logic for computing and signal processing over the past decade. System advantages that the optical schemes promise over the conventional electrical schemes including but not limited to: significant reduction of gate latency, ultra-low energy consumption, and simplified layout architecture. For example, adding is the fundamental operation for computation. However, though it has a concise layout, a conventional electrical carry-ripple adder will be too slow for many-bit addition due to excessive carry propagation delay. Thus, modern microprocessors all have to use much more complicated structures such as parallel prefix adders to obtain satisfactory performance while they come with inevitable power penalty. The greater the number of bits, the more complex and power hungry the adder will become. In this paper, we propose a silicon photonics based architecture of carry-ripple adder, which utilizes the particular merit of light that interference, for future high-speed and low-power consumption optical computing. In our proposed carry-ripple adder, the critical path will be built with optical switch and optical signals will be modulated to carry information. As all of the optical switches could operate simultaneously, the unpleasant accumulated gate latency from the electrical approach will be removed. Silicon photonic based optical switches are promising candidates to implement the adder due to their compact sizes, which could significantly reduce the capacitances and energy-consumption.
Due to the bottleneck in the continuation of Moore’s law as well as the drastically increasing trend of bandwidth, silicon photonics has emerged as the most promising candidate for implementing next-generation communication networks with ultralow power and ultrahigh speed. Recently, optical computing in integrated photonics, which outperforms electrical counterparts both in power consumption and bandwidth, has attracted a renewed interest due to the accessibility and maturity of ultracompact passive and active integrated components. However, up to now, most of relevant research about optical computing still focus on the realization of fundamental logic gates, not even close to feasible large-scale computing system. In this paper, we demonstrate a high-speed ripple-carry electro-optic full adder using micro-resonators. This approach adopts photons instead of electrons to realize logic operations as well as transfer carry signals from one bit to the next, while all the control signals of operands are applied simultaneously at and within every clock cycle. Thus, the severe latency issue that accumulates as the size of full adder increases can be circumvented, allowing for the improvement in computing speed. This approach also outperforms the conventional electrical counterpart in terms of power consumption due to the relatively smaller propagation loss and switching energy.
With the increasing demand for information, integrated silicon photonics technology has been highly valued. Among
them, silicon on insulator (SOI) has advantages of low cost, process maturity, and IC technology compatible, making
it to be one of the most competitive integrated optoelectronic platforms. However, due to its highly
polarization-dependent performance, polarization-selective devices are essential on SOI platform. In this paper, we
analyze the critical guiding condition of SOI waveguide as well as the hybrid plasmonic waveguide (HPW). Based
on the different critical guiding condition for both polarizations, we propose several polarization-selective devices on
SOI platform, including polarizer, polarization beam splitter (PBS) and polarization rotator. In this paper, an
ultracompact and broadband TE-pass polarizer based on HPW is proposed. In addition, an asymmetrical coupling
based polarization beam splitter is designed. Simulation results show that the designed devices have excellent optical
properties, and the sizes of the devices achieve a great breakthrough.
Capacitive gate transient noise has been problematic for the high-speed single photon avalanche photodiode (SPAD), especially when the operating frequency extends to the gigahertz level. We proposed an electro-optic modulator based gate transient noise suppression method for sine-wave gated InGaAs/InP SPAD. With the modulator, gate transient is up-converted to its higher-order harmonics that can be easily removed by low pass filtering. The proposed method enables online tuning of the operating rate without modification of the hardware setup. At 250 K, detection efficiency of 14.7% was obtained with 4.8×10−6 per gate dark count and 3.6% after-pulse probabilities for 1550-nm optical signal under 1-GHz gating frequency. Experimental results have shown that the performance of the detector can be maintained within a designated frequency range from 0.97 to 1.03 GHz, which is quite suitable for practical high-speed SPAD applications operated around the gigahertz level.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.