Optical Neuron Network Using Waveguide Coupling for Nonlinear Processing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional optical processing systems are limited in handling nonlinear processing operations, which are typically operated by electronic processing or high-energy laser units, and lack efficient all-optical neuron network configurations capable of handling both data transmission and processing using optical manipulations.

Innovation Solution

The development of an artificial optical neuron network utilizing multi-mode and multi-core optical fibers, along with free-space propagation, enables fully operated all-optical neuron networks with controlled couplings, processing operations, and training processes, including modal mixing, gain application, and spatial/temporal signal portion mixing to adjust weights within the network.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional optical processing systems are used, then data transmission can be performed, but nonlinear processing operations cannot be handled efficiently without electronic conversion

Engineering Contradiction:

Improvenonlinear processing capabilityVSAvoidsystem architecture

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent replaces electronic processing mechanisms with optical mechanisms by using nonlinear optical effects (such as Kerr effect, four-wave mixing) in optical fibers to perform nonlinear operations directly on optical signals, eliminating the need for OEO conversion and achieving all-optical processing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent employs composite optical structures including multi-core fibers, multi-mode fibers, and photonic crystal fibers that combine different material properties to enable both linear data transmission and nonlinear processing operations within the same optical medium

Inventive Principle:

Principle #40Composite materials

2Loss of energy

If electronic-optical hybrid processing is used, then certain processing operations can be performed, but energy loss occurs during conversion between electronic and optical signals

Engineering Contradiction:

Improveenergy lossVSAvoidprocessing operation capability

Core Design Contradiction:

Loss of energyVSAdaptability or versatility

Solution Approach 1:

The patent substitutes electronic processing with optical processing throughout the entire signal path, using optical amplifiers, optical modulators, and nonlinear optical devices to perform all processing operations optically, thereby eliminating energy loss associated with OEO conversions while maintaining full processing capability

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent maintains continuous optical signal flow through the system without interruption for electronic conversion, using optical buffering, optical switching, and all-optical amplification to ensure uninterrupted optical processing and transmission

Inventive Principle:

Principle #20Continuity of useful action

3Speed

If electronic-optical hybrid processing is used, then data can be processed, but transmission speed is reduced due to conversion requirements

Engineering Contradiction:

Improvetransmission speedVSAvoidconversion requirements

Core Design Contradiction:

SpeedVSDevice complexity

Solution Approach 1:

The patent replaces all electronic conversion mechanisms with optical mechanisms, using direct optical modulation, optical switching, and optical signal processing to eliminate conversion delays and achieve transmission speeds limited only by the speed of light in the optical medium

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent performs preliminary optical processing operations (such as pre-amplification, pre-modulation, and wavelength conversion) before signal transmission to minimize the need for intermediate conversions and maintain high speed throughout the signal path

Inventive Principle:

Principle #10Preliminary action

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This solution allows for efficient all-optical neuron networks that perform nonlinear operations without the need for electronic signal conversion, enhancing processing speed and reducing energy consumption by leveraging optical manipulations for data transmission and processing.

Implementation Method 1

a modal mixing unit, such as multimode optical fiber, configured for receiving input light and applying selected mixing to light components of two or more modes within the input light

Methodology Applied

Scientific EffectModal mixing: Diffusion

Implementation Method 2

having at least a potion thereof impregnated with gain medium and comprising predetermined gain medium configured for emitting light at predetermined first wavelength range in response to pumping light of second wavelength range

Methodology Applied

Scientific EffectStimulated emission: Laser

Data Source

PatentUS20240078419A1Optical neuron unit and network of the same

Publication Date: 2024.03.07 COGNIFIBER LTD

AI summary

An artificial neuron network and corresponding neuron units are described and corresponding neuron units. The neuron network comprises a plurality of two or more layers of artificial neuron units. The layers of artificial neuron units are configured for communicating between them via an arrangement of two or more optical waveguide (optical fibers). The arrangement of two or more optical waveguides are configured with predetermined coupling between the two or more waveguides, thereby providing cross communication between neuron units of said two or more layers.