Mott Insulator Artificial Neuron for Neuromorphic Integration

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current artificial neurons are complex and energy-intensive, requiring multiple electronic components and being difficult to miniaturize, which limits their integration and efficiency in neuromorphic electronic circuits.

Innovation Solution

A single-component artificial neuron made of a Mott insulator electric dipole with conductive electrodes, capable of performing the three functions of integration, leakage, and firing, simplifying production and enhancing integration potential.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If conventional multi-component artificial neurons are used, then the neuron can perform integration, leakage, and firing functions, but the device complexity and energy consumption increase significantly

Engineering Contradiction:

Improvenumber of electronic componentsVSAvoidenergy consumption

Core Design Contradiction:

Device complexityVSUse of energy by moving object

Solution Approach 1:

The patent merges the functions of integration, leakage, and firing into a single artificial neuron component, eliminating the need for multiple separate electronic components. This consolidation directly reduces device complexity and the associated energy consumption while maintaining all necessary neuronal functions.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The artificial neuron is designed as a universal component that simultaneously performs multiple functions (integration, leakage, and firing) that were previously requiring separate components. This multi-functionality approach reduces the overall number of components needed in the neuromorphic circuit.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Device complexity

If conventional artificial neurons with multiple components are used, then the neuron functions are achieved, but the miniaturization and integration potential are limited

Engineering Contradiction:

Improvenumber of electronic componentsVSAvoidintegration potential

Core Design Contradiction:

Device complexityVSVolume of moving object

Solution Approach 1:

By combining all neuronal functions into a single component, the patent enables significant miniaturization. The reduced component count directly translates to smaller overall device volume and higher integration potential, allowing more neurons to be packed into a given area.

Inventive Principle:

Principle #5Merging (Combining)

3Ease of manufacture

If conventional artificial neurons are used, then the neuron can process signals, but the production complexity increases

Engineering Contradiction:

Improveproduction simplicityVSAvoidnumber of electronic components

Core Design Contradiction:

Ease of manufactureVSDevice complexity

Solution Approach 1:

The consolidation of multiple components into a single artificial neuron simplifies the manufacturing process. Fewer components mean fewer assembly steps, reduced placement complexity, and easier production scaling, directly improving ease of manufacture.

Inventive Principle:

Principle #5Merging (Combining)

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

The single-component artificial neuron achieves efficient integration and firing functions with reduced energy consumption and increased integration potential, simplifying the production of artificial neural networks.

Implementation Method 1

exhibiting the phenomenon of electrically induced volatile resistive transition

Methodology Applied

Scientific EffectMott insulator resistive transition:

Data Source

PatentEP3138050B1Single-component artificial neuron made from mott insulators, network of artificial neurons, and method for making said artificial neurons

Publication Date: 2021.05.05 CENT NAT DE LA RECH SCI (C N R S)

Figure 1~2
Figure 3~4B
Figure 5A~6

AI summary

The invention relates to an artificial neuron (30) consisting of a single-component electric dipole comprising a single material (31) which belongs to the class of Mott insulators and is connected to two electric electrodes (32, 33).