CN115202127B - Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator - Google Patents

Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator

Info

Publication number
CN115202127B
CN115202127B CN202210858513.0A CN202210858513A CN115202127B CN 115202127 B CN115202127 B CN 115202127B CN 202210858513 A CN202210858513 A CN 202210858513A CN 115202127 B CN115202127 B CN 115202127B
Authority
CN
China
Prior art keywords
nonlinear
silicon
power
output
ring resonator
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202210858513.0A
Other languages
Chinese (zh)
Other versions
CN115202127A (en
Inventor
王曰海
刘笑之
杨建义
余辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shaoxing Research Institute Of Zhejiang University
Original Assignee
Shaoxing Research Institute Of Zhejiang University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shaoxing Research Institute Of Zhejiang University filed Critical Shaoxing Research Institute Of Zhejiang University
Priority to CN202210858513.0A priority Critical patent/CN115202127B/en
Publication of CN115202127A publication Critical patent/CN115202127A/en
Application granted granted Critical
Publication of CN115202127B publication Critical patent/CN115202127B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G02—OPTICS
    • G02F—OPTICAL DEVICES OR ARRANGEMENTS FOR THE CONTROL OF LIGHT BY MODIFICATION OF THE OPTICAL PROPERTIES OF THE MEDIA OF THE ELEMENTS INVOLVED THEREIN; NON-LINEAR OPTICS; FREQUENCY-CHANGING OF LIGHT; OPTICAL LOGIC ELEMENTS; OPTICAL ANALOGUE/DIGITAL CONVERTERS
    • G02F1/00—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics
    • G02F1/35—Non-linear optics
    • G02F1/3501—Constructional details or arrangements of non-linear optical devices, e.g. shape of non-linear crystals
    • G—PHYSICS
    • G02—OPTICS
    • G02F—OPTICAL DEVICES OR ARRANGEMENTS FOR THE CONTROL OF LIGHT BY MODIFICATION OF THE OPTICAL PROPERTIES OF THE MEDIA OF THE ELEMENTS INVOLVED THEREIN; NON-LINEAR OPTICS; FREQUENCY-CHANGING OF LIGHT; OPTICAL LOGIC ELEMENTS; OPTICAL ANALOGUE/DIGITAL CONVERTERS
    • G02F1/00—Devices or arrangements for the control of the intensity, colour, phase, polarisation or direction of light arriving from an independent light source, e.g. switching, gating or modulating; Non-linear optics
    • G02F1/35—Non-linear optics
    • G02F1/355—Non-linear optics characterised by the materials used
    • G02F1/3556—Semiconductor materials, e.g. quantum wells
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/067—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using optical means

Landscapes

  • Physics & Mathematics (AREA)
  • Nonlinear Science (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Optics & Photonics (AREA)
  • Theoretical Computer Science (AREA)
  • Neurology (AREA)
  • Software Systems (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Chemical & Material Sciences (AREA)
  • Crystallography & Structural Chemistry (AREA)
  • Optical Modulation, Optical Deflection, Nonlinear Optics, Optical Demodulation, Optical Logic Elements (AREA)

Abstract

本发明公开了一种基于硅基微环谐振器的全光非线性激活函数的实现方法，本发明通过上下话路的微环谐振器，基于不同的光波输出端口得到不同谐振图谱，通过激活非线性硅基微环谐振器谐振波长和未激活非线性硅基微环谐振器谐振波长分别对不同谐振图谱进行区域划分，在不同区域内的通过输入功率和输出功率分别构建radial bias非线性函数，Gauss非线性函数，Asymptotic非线性函数和Sigmoid非线性函数。实现具有超快时间响应和超低能耗的光学神经网络。

This invention discloses a method for implementing an all-optical nonlinear activation function based on a silicon-based microring resonator. The method utilizes microring resonators with uplink and downlink paths to obtain different resonance spectra based on different optical output ports. By dividing the resonance spectra into regions using the resonant wavelengths of both the activated and unactivated nonlinear silicon-based microring resonators, and constructing radial bias, Gaussian, asymptotic, and sigmoid nonlinear functions within each region using input and output power, respectively, this method achieves an optical neural network with ultrafast time response and ultra-low power consumption.

Description

Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator
Technical Field
The invention belongs to nonlinear optics, and particularly relates to a realization method of an all-optical nonlinear activation function based on a silicon-based micro-ring resonator.
Background
In emerging internet of things systems, big data analysis requires powerful computing power and higher energy efficiency. Device integration in electronic circuits such as microprocessor chips has steadily evolved at the speed of moore's law due to limitations imposed by Dennard scaling techniques. As moore's law tends to end, the clock frequency of the processor tends to settle after 2004. The multiple processors then help maintain a steady increase in throughput through parallel computing. However, according to the law of armdar, the speed of parallel computation is limited, and parallel computation cannot solve all problems.
As electronic devices face the above problems, the development of optical neural networks is rapid. The high bandwidth, low latency processing capability of photonics combined with the distributed processing of artificial neural networks is uniquely prepared for real-time processing that has not been possible with previous electronic devices. The basic unit of the photonic neural network consists of linear matrix vector multiplication and nonlinear activation. Nonlinear activation in neuromorphic photonic hardware may be achieved by photoelectric conversion or all-optical. Whereas most advanced optoelectronic hardware still has bandwidth and speed limitations in optoelectronic (OEO) conversion, optical nonlinearity has advantages of high speed and high integration, so optical nonlinearity has higher value.
The main structures in the design of most of the existing optical neural networks of on-chip integrated waveguides and other on-chip optical computing platforms are beam splitting waveguide units and cascaded Mach-Zehnder interferometers, mach-zehnderinterferometers (mzi), which are only suitable for linear computing. Because superposition of linear calculation is still linear calculation, the calculation result of the design can be equal to one-time matrix multiplication operation no matter how the total layers are, the parameter range is limited, and the requirement of the neural network for fitting the data cannot be met. Such optical neural networks rely on further nonlinear calculations in electronic circuits during subsequent information processing, and cannot integrate the functionality of a complete neural network on an optical platform. In addition, an inherent compromise is often presented between ultrafast response time and huge nonlinearity, so that a larger nonlinear coefficient can only usually come at the cost of slower response time. Therefore, efficient and feasible nonlinear calculation is difficult to realize in the transplanting process of a large-scale optical hardware platform, so that an optical nonlinear activation layer only stays in a theoretical concept and cannot realize practical application.
The reason why nonlinear calculation is difficult on an optical platform is that the total light nonlinear effect of the material is weak, the nonlinear material with enough strength is lacking, and the strong nonlinear effect is difficult to realize in an on-chip integrated device, while the gst material which can be used for nonlinear calculation has nonvolatile characteristic and is not suitable for high-efficiency calculation of quick response. Therefore, the real introduction of an all-optical nonlinear active layer is still a urgent problem to be solved for the realization of an optical neural network with ultra-fast time response and ultra-low energy consumption for a hardware platform.
Disclosure of Invention
The invention provides a method for realizing an all-optical nonlinear activation function based on a silicon-based micro-ring resonator, which can realize the all-optical nonlinear activation function simply by regulating and controlling the wavelength and the optical power range of input light waves.
A realization method of an all-optical nonlinear activation function based on a silicon-based micro-ring resonator, wherein the silicon-based micro-ring resonator is a micro-ring resonator with an upper voice channel and a lower voice channel and comprises an input end, a Drop end and a Through end, when the input power is W 1, the nonlinearity of the silicon-based micro-ring resonator cannot be excited, when the input power is increased from W 1 to W 2, the nonlinearity of the silicon-based micro-ring resonator is excited, when an optical wave is output from the Through end, the input power is increased from W 1 to W 2, and the input power and the corresponding output power are fitted to obtain a first nonlinear function;
When the input power is increased from W 1 to W 3, the nonlinearity of the silicon-based micro-ring resonator is further excited, when the light wave is output from the Drop end, the input power is increased from W 1 to W 3, and a second nonlinear function can be obtained by fitting the input power with the corresponding output power.
When the input power is increased to W 2, the micro-ring resonator is excited in a nonlinear mode, the frequency spectrum shifts, the resonance wavelength of the silicon-based micro-ring resonator is changed from lambda 1 to lambda 2, and when the wavelength of the light wave is located in a first interval near lambda 1, the input power and the output power of the light wave show GeLU nonlinear functions.
When the wavelength of the lightwave is within a second interval between λ 1 and λ 2, the input power and output power of the lightwave exhibit a radial bias nonlinear function.
When the wavelength of the light wave is within the third interval around lambda 2, the input power and output power of the light wave exhibit Gauss nonlinear functions.
When the input power is increased to W 3, the nonlinearity of the micro-ring resonator is excited, the frequency spectrum shifts, the resonance wavelength of the silicon-based micro-ring resonator is changed from lambda 1 to lambda 3, and when the wavelength of the light wave is in a fourth wavelength interval near lambda 1, the input power and the output power of the light wave show a Asymptotic nonlinear function.
When the wavelength of the lightwave is within the fifth wavelength interval between λ 1 and λ 3, the input power and output power of the lightwave exhibit Sigmoid nonlinear functions.
When the wavelength of the light wave is greater than lambda 3, the input power and output power of the light wave exhibit a power function.
The silicon-based material in the silicon-based micro-ring resonator is silicon material, and nonlinear loss occurs to the silicon material through Kerr effect, two-photon absorption effect, free carrier dispersion or free carrier absorption effect.
When W 1＝0.1mW,W2＝1mW,W3 = 2mW;
Lambda 1＝1547.7623nm、λ2＝1547.7806nm、λ3 = 1547.7968 nm;
The first range of the Through end is lambda 1-0.0023nm,λ1 +0.0047nm, the second range lambda 1+0.0077nm,λ2 -0.0026nm, and the third range lambda 2,λ2 +0.0014nm.
The fourth interval range of the Drop end is lambda 1-0.0423nm,λ1 +0.0047nm, and the fifth interval lambda 1+0.0177nm,λ3 -0.0092nm.
Compared with the prior art, the invention has the beneficial effects that:
According to the invention, different resonance patterns are obtained based on different light wave output ports through micro-ring resonators of upper and lower voice channels, different resonance patterns are respectively divided by activating the resonance wavelength of the nonlinear silicon-based micro-ring resonator and the resonance wavelength of the non-activated nonlinear silicon-based micro-ring resonator, and a radial bias nonlinear function, a Gauss nonlinear function, a Asymptotic nonlinear function and a Sigmoid nonlinear function are respectively constructed in different areas through input power and output power. An optical neural network with ultra-fast time response and ultra-low energy consumption is realized.
Drawings
In order to more clearly illustrate the prior art and the present invention, the drawings used in the description of the prior art and the embodiments of the present invention will be briefly described. It will be apparent to those skilled in the art that the drawings in the following description are merely exemplary and that other drawings may be derived from the drawings provided without the inventive effort to those skilled in the art.
The structures, proportions, sizes, etc. shown in the present specification are shown only for the purposes of illustration and description, and are not intended to limit the scope of the invention, which is defined by the claims, for example, modifications, variations in proportions, or otherwise, used in the practice of the invention, which are particularly adapted to specific environments without departing from the spirit and scope of the invention.
Fig. 1 is a block diagram of an upper and lower session micro-ring according to an embodiment.
FIG. 2 is a graph showing the frequency spectrum of the Through end of a silicon-based microring at 0.1mW (non-excited non-linearity) and 1mW (excited non-linearity) provided in the embodiment.
Fig. 3 is three types of nonlinear graphs obtained when the output port provided in the embodiment is the Through end.
Fig. 4 is a graph of a nonlinear function obtained by three types of nonlinear curve fitting, where the output port provided in the embodiment is a Through end.
FIG. 5 is a graph of a silica-based microring Drop-end spectrum at 0.1mW (unexcited nonlinearity) and 2mW (excited nonlinearity) provided in an embodiment.
Fig. 6 is three types of nonlinear graphs obtained by using Drop as an output port according to the embodiment.
Fig. 7 is a graph of a nonlinear function obtained by three types of nonlinear curve fitting, where the output port provided in the embodiment is Drop.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the detailed description is presented by way of example only and is not intended to limit the scope of the invention.
The nonlinear function is extracted by the upper and lower voice channel micro-loop structure diagram shown in fig. 1.
The silicon-based material of the silicon-based micro-ring resonator is a silicon material, and according to fig. 2, when light is output from the Through end, micro-ring transmission spectra are obtained at 0.1mW (non-linear unexcited) and 1mW (non-linear excited), and the resonance wavelengths are respectively lambda 1 = 1547.7623nm and lambda 2 = 1547.7806nm. The visible spectrum may undergo a red shift due to the presence of third order nonlinear effects. The spectrum is divided into three areas, wherein the linear resonance wavelength of the area I is located near lambda 1, the wavelength range is 1547.760nm-1547.767nm, the area II is between lambda 1、λ2, the wavelength range is 1547.770nm-1547.778nm, the area III is on the right side of lambda 2, and the wavelength range is 1547.7806nm-1547.782nm. Nonlinear functions may also be obtained outside the region.
According to fig. 3, when the port is selected at the Through end, the input power is gradually increased from 0.1mW to 1mW at a fixed wavelength, and the output power at different input powers is recorded, and each point is connected to obtain the input-output nonlinear curve. When three wavelengths are selected from the three regions, respectively, three classes GELU, radial bias, gauss of nonlinear functions are obtained.
As can be obtained from fig. 4, the curves obtained from the Through end are subjected to nonlinear fitting, and the formulas GELU, the radial bias and the Gauss are respectively as follows:
y=(0.9991x-0.0409)/(1.16687+e^(-4.08249x+1.38812)),R2=0.9999;
y=-0.02862+1.3292x-6.11838x^2+9.78334x^3-4.43319x^4,R2=0.99504;
y=-0.26616+0.64275/(w(0.5π)^0.5)*e^(-2(x-0.55709)^2/w^2),w=0.93872,R2=0.99766。
Where R 2 is a linear regression coefficient, a closer to 1 indicates a higher degree of curve fitting. It can be seen that the nonlinear function obtained from the micro-ring matches GELU, radial bias, gauss very well.
As can be seen from fig. 5, the microring transmission spectra at 0.1mW (non-linear unexcited) and 2mW (non-linear excited) when light is output from the Drop end have resonance wavelengths of λ 1＝1547.7623nm,λ3 = 1547.7968nm, respectively. The spectrum is divided into two regions, region I near lambda 1, wavelength range 1547.72nm-1547.770nm, region II between lambda 1、λ3, and wavelength range. Non-linear functions are also available outside the region but the shape cannot match the existing function type. The range interval is mainly obtained by debugging.
As can be obtained from fig. 6, when the port is selected at the Drop end, the input power is gradually increased from 0.1mW to 2mW, the output power at different input powers is recorded, and each point is connected to obtain the input-output nonlinear curve. When a wavelength is selected from each of the two regions, a Sigmoid, asymptotic function is ultimately obtained. The power function is available to the right of lambda 3 but is not available for nonlinear activation and is therefore not considered.
As can be taken from fig. 7, the curve fitting from Drop end, sigmoid, asymptotic formula is:
y=0.75769-0.759*0.26093x,R2=0.99968;
y=1.57944/(1+e^(-3.41955(x-0.90652))),R2=0.99751。
Where R 2 is a linear regression coefficient, a closer to 1 indicates a higher degree of curve fitting. It can be seen that the functions obtained from different regions of Drop have a higher degree of matching with Sigmoid, asymptotic, providing a higher probability for use in neural networks.

Claims (2)

1. The realization method of the all-optical nonlinear activation function based on the silicon-based micro-ring resonator is characterized in that the silicon-based micro-ring resonator is a micro-ring resonator with an upper voice channel and a lower voice channel, and comprises an input end, a Drop end and a Through end;
When the output port is selected at the Through end, the wavelength is fixed, the input power is gradually increased from 0.1mW to 1mW, the output power under different input powers is recorded, each point is connected to obtain an input-output nonlinear curve, and when three wavelengths are selected from three areas respectively, nonlinear functions of three categories GELU, radial bias and Gauss can be obtained;
The light is output from the Through end, microring transmission spectra are respectively obtained under the conditions of 0.1mW, namely the power when nonlinearity is not excited, 1mW, namely the power when nonlinearity is excited, the resonant wavelengths are λ1= 1547.7623nm and λ2= 1547.7806nm, the visible spectrum can be subjected to red shift due to the existence of a third-order nonlinear effect, the spectrum is divided into three areas, the linear resonant wavelength of the area I is positioned near λ1, the wavelength range is 1547.760nm-1547.767nm, the area II is between λ1 and λ2, the wavelength range is 1547.770nm-1547.778nm, the area III is positioned on the right side of λ2, and the wavelength range is 1547.7806nm-1547.782nm;
When the output port is selected at the Through end, the wavelength is fixed, the input power is gradually increased from 0.1mW to 1mW, the output power under different input powers is recorded, each point is connected to obtain an input-output nonlinear curve, and when three wavelengths are selected from three areas respectively, nonlinear functions of three categories GELU, radial bias and Gauss can be obtained;
Or when light is output from a Drop end, respectively carrying out microring transmission spectrograms under the conditions of 0.1mW, namely non-linear unexcited power and 2mW, namely non-linear excited power, wherein the resonant wavelength is lambda 1＝1547.7623nm,λ3 = 1547.7968nm, dividing the spectrum into two areas, wherein the area I is near lambda 1, the wavelength range is 1547.72nm-1547.770nm, the area II is between lambda 1、λ3, when the port is selected at the Drop end, gradually increasing the input power from 0.1mW to 2mW, recording the output power under different input powers, connecting the points to obtain an input-output non-linear curve, and when one wavelength is selected from the two areas, obtaining a Sigmoid, asymptotic function, and when the right side of lambda 3 is available as a power function but not available for non-linear activation.
2. The method for realizing the all-optical nonlinear activation function based on the silicon-based micro-ring resonator according to claim 1, wherein the silicon-based material in the silicon-based micro-ring resonator is a silicon material, and the silicon material is subjected to nonlinear loss through a Kerr effect, a two-photon absorption effect, free carrier dispersion or a free carrier absorption effect.
CN202210858513.0A 2022-07-20 2022-07-20 Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator Active CN115202127B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210858513.0A CN115202127B (en) 2022-07-20 2022-07-20 Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210858513.0A CN115202127B (en) 2022-07-20 2022-07-20 Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator

Publications (2)

Publication Number Publication Date
CN115202127A CN115202127A (en) 2022-10-18
CN115202127B true CN115202127B (en) 2026-01-13

Family

ID=83581326

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210858513.0A Active CN115202127B (en) 2022-07-20 2022-07-20 Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator

Country Status (1)

Country Link
CN (1) CN115202127B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115936087A (en) * 2022-12-27 2023-04-07 北京工业大学 An Optical Neuron Structure Based on On-Chip Microring Resonator
CN118192010B (en) * 2024-02-01 2024-08-13 重庆邮电大学 Silicon-based light-controlled microring resonator for in-situ training of optical neural networks

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113673677A (en) * 2021-08-18 2021-11-19 山东云海国创云计算装备产业创新中心有限公司 Method, equipment and medium for realizing nonlinear activation function RELU

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8760661B2 (en) * 2011-05-03 2014-06-24 Howard Hughes Medical Institute Method and apparatus for compressing ultrashort optical pulses
CN103744251A (en) * 2013-12-19 2014-04-23 杭州电子科技大学 All-optical trigger based on nonlinear micro-ring resonant cavity
CN113961035B (en) * 2021-10-15 2024-12-10 北京理工大学 High-efficiency reconfigurable all-optical neural network computing chip architecture for deep learning
CN114218834A (en) * 2021-12-16 2022-03-22 西安石油大学 Optical frequency comb performance analysis method and system based on machine learning
CN114548381B (en) * 2022-04-25 2022-08-05 苏州浪潮智能科技有限公司 Activation function construction method, device, equipment and medium of all-optical neural network

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113673677A (en) * 2021-08-18 2021-11-19 山东云海国创云计算装备产业创新中心有限公司 Method, equipment and medium for realizing nonlinear activation function RELU

Also Published As

Publication number Publication date
CN115202127A (en) 2022-10-18

Similar Documents

Publication Publication Date Title
CN114358271B (en) Time-wavelength interweaving photon neural network convolution acceleration chip
Donati et al. Microring resonators with external optical feedback for time delay reservoir computing
CN113961035B (en) High-efficiency reconfigurable all-optical neural network computing chip architecture for deep learning
Li et al. All-optical nonlinear activation function based on germanium silicon hybrid asymmetric coupler
Ling et al. On-chip optical matrix-vector multiplier based on mode division multiplexing
CN113890620B (en) Photonic neural network on silicon substrate based on tunable filter and its modulation method
CN115202127B (en) Method for realizing all-optical nonlinear activation function based on silicon-based micro-ring resonator
Hou et al. Programmable optical filter in thin-film lithium niobate with simultaneous tunability of extinction ratio and wavelength
Bharti et al. Design of micro-ring resonator based all optical universal reconfigurable logic circuit
Saharia et al. Proposed all-optical read-only memory element employing Si3N4 based optical microring resonator
Sasikala et al. Performance analysis of multilogic all-optical structure based on nonlinear signal processing in SOA
CN116774351A (en) Lithium niobate-based optical power distributor with arbitrary proportion and design method
Tripathi Evaluating RSOA performance with optical logic gates at 100 Gbps data rate
Hossain et al. Numerical analysis of all-optical binary to gray code converter using silicon microring resonator
KR100418654B1 (en) All-Optical XOR Gate by using Semiconductor Optical Amplifier
Kaur et al. All-optical comparator using logic operations based on nonlinear properties of semiconductor optical amplifier
CN117032392A (en) Silicon-based polarization multiplexing integrated optical computing chip
Kaur et al. SOA-MZI based 4× 4 interconnected crossbar photonic wavelength switching for datacenter load balancing
Mukherjee et al. All-optical Feynman gate using reflective semiconductor optical amplifiers and binary to gray code converter
Mourgias-Alexandris et al. Optical thresholding device with a sigmoidal transfer function
LeGrange et al. Cascaded all-optical operations in a hybrid integrated 80-Gb/s logic circuit
JP3805286B2 (en) Method and apparatus for realizing all-optical half-adder using semiconductor optical amplifier
CN118033928A (en) Integrated ultrashort pulse generator based on thin film lithium niobate
Jin et al. Photonic integrator for A/D conversion
US20250028949A1 (en) Optical neural network with gain from parity time optical couplers

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant