Loading...
News Article

VCSEL used in novel photonic neural network

News

Strathclyde University team achieves high-accuracy classification using a VCSEL and new training method

Photonic neural network systems, which are fast and energy efficient, are helpful for dealing with large amounts of data. To advance photonic brain-like computing technologies, a group of researchers at the University of Strathclyde combined a spike-based neural network with a VCSEL that exhibits spiking neuronal behaviours.

Recently, they presented high-performance photonic spiking neural network operation with lower training requirements and introduced a novel training scheme for getting better results. This research was published Aug. 29 in the journal Intelligent Computing.

The spiking neural network proposed by the authors is based on a hardware-friendly photonic system consisting of just one VCSEL. This proposal represents an improvement on an earlier study by the same authors in which they combined reservoir computing, a powerful method for building photonic neural networks and tackling complex tasks, with a neuromorphic spiking neuron built using a laser of the same type.

In the current paper, the authors accomplished a more challenging classification task and applied an alternative training scheme to improve the speed and efficiency of training while reducing training requirements.

The classification task they tackled is a complex, multivariate and nonlinear problem involving 500 features per data point, and is based on the artificial dataset MADELON. To create the spiking neural network, the authors used an experimental setup that combines the nonlinear spiking dynamics of the laser with an architecture inspired by reservoir computing.

In this architecture, the input data is time-multiplexed. Each time slot represents a virtual neuron in the neural network. The input data is injected into and processed by the laser, and the output is interpreted as a binary node output, either spiking or non-spiking, depending on whether the input data exceeds a certain threshold.

The authors successfully demonstrated the computational power of the photonic spiking neural network with both a traditional least-squares regression training method and their newly proposed “significance” training method. The latter assigns binary weights to nodes based on their overall usefulness and significance.

They report that both methods achieved excellent classification accuracies of over 94 percent, surpassing the benchmark performance in a fraction of the processing time. The accuracies achieved using the new method were 94.4 percent and 95.7 percent, higher than those achieved by the traditional method.

The authors believe that this study could create new possibilities for photonics-based processing systems that operate entirely on optical hardware, enabling them to tackle highly complex tasks with high accuracy, high speed, and energy-efficient operation.

Reference

'Photonic Spiking Neural Networks with Highly Efficient Training Protocols for Ultrafast Neuromorphic Computing Systems' by Dafydd Owen-Newns et al; Intelligent Computing 29 Aug (2023)

SPONSOR MESSAGE

Secure Your Hydrogen Supply

A study supply of high-purity hydrogen is critical to semiconductor fabrication. Supply chain interruptions are challenging manufacturers, leading to production slowdowns and stoppages. On-site hydrogen generation offers a scalable alternative for new and existing fabs, freeing the operator from dependence on delivered gas.

Plant managers understand the critical role that hydrogen plays in semiconductor fabrication. That important job includes crystal growth, carrier gas, wafer annealing, and in the emerging Extreme UV Lithography (EUV) that will enable new generations of devices. As the vast need for semiconductors grows across all sectors of world economies, so does the need for high-purity hydrogen.

Take control with Nel on-site hydrogen generation.

Read more
QD company Quantum Science expands into new facility
Innoscience files lawsuit against Infineon
Riber revenues up 5% to €41.2m
Forvia Hella to use CoolSiC for next generation charging
Photon Design to exhibit QD simulation tool
Ortel transfers CW laser fabrication to Canada
Luminus adds red and blue multi-mode Lasers
PseudolithIC raises $6M for heterogeneous chiplet tech
Mesa sidewall design improves HV DUV LEDs
IQE revenue to exceed expectations
'Game-changing' VCSEL system targets clinical imaging
German start-up secures finance for SiC processing tech
Macom signs preliminaries for CHIPS Act funding
IQE and Quintessent partner on QD lasers for AI
EU funds perovskite tandems for fuel-free space propulsion
EU to invest €3m in GeSi quantum project
Transforming the current density of AlN Schottky barrier diodes
Turbocharging the GaN MOSFET with a HfO₂ gate
Wolfspeed launches Gen 4 SiC MOSFET technology
Report predicts high growth for UK's North East
Element Six unveils Cu-diamond composite
SemiQ launches hi-rel 1700V SiC MOSFETs
Lynred to exhibit Eyesential SWIR sensor for machine vision
Thorlabs buys VCSEL firm Praevium Research
Advancing tuneable InP lasers on a heterogeneous platform
P-GaN gate HEMTs have record threshold voltage
Guerrilla RF releases GaN power amplifier dice
Narrow-linewidth DFB lasers now at 405 and 488nm
Researchers develop tech for future fast-charging stations
Vermont GaN Tech Hub awarded nearly $24M
Onsemi completes buy-out of Qorvo SiC JFET business
Quantum Science announces Innovate UK funding
×
Search the news archive

To close this popup you can press escape or click the close icon.
Logo
x
Logo
×
Register - Step 1

You may choose to subscribe to the Compound Semiconductor Magazine, the Compound Semiconductor Newsletter, or both. You may also request additional information if required, before submitting your application.


Please subscribe me to:

 

You chose the industry type of "Other"

Please enter the industry that you work in:
Please enter the industry that you work in:
 
x
Adblocker Detected
Please consider unblocking adverts on this website