3D Nonlinear Element for Compact Machine Learning Systems

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional neural network and machine learning systems face limitations in compactness due to the two-dimensional arrangement of nonlinear units, which affects their efficiency and applicability in compact form factors.

Innovation Solution

A three-dimensional electric element comprising four or more nonlinear units with nonlinear current-voltage characteristics connected by an electric conductor, arranged in a three-dimensional manner, along with a manufacturing method using a dispersion liquid and curable resin to create a compact machine learning system.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Volume of moving object

If nonlinear units are arranged in a two-dimensional manner, then the system structure is simple and easy to manufacture, but the system compactness is limited

Engineering Contradiction:
Improvesystem compactnessVSAvoidarrangement structure
Core Design Contradiction:
Volume of moving objectVSDevice complexity

Solution Approach 1:

The patent transitions from a two-dimensional planar arrangement of nonlinear units to a three-dimensional spatial arrangement. This dimensionality change allows the system to achieve higher compactness by utilizing vertical stacking and multi-layer configurations, thereby reducing the overall footprint and volume occupation of the neural network system while maintaining manufacturing feasibility through adapted fabrication processes.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Volume of moving object

If conventional functional molecular elements are used, then the system can be made compact, but there is a limit to the compactness achievement

Engineering Contradiction:
Improvesystem compactnessVSAvoidneural network performance
Core Design Contradiction:
Volume of moving objectVSReliability

Solution Approach 1:

The patent applies local quality by optimizing the arrangement density and connectivity of nonlinear units in specific three-dimensional regions. Different areas of the 3D structure can have varying densities and configurations tailored to specific functional requirements, allowing high compactness in critical regions while maintaining overall system reliability and performance through localized optimization rather than uniform compression.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS20230376737A13-dimensional electrical element, machine learning system comprising same, and methods for manufacturing said element and said system
Publication Date: 2023.11.23 NAT UNIV CORP KYUSHU INST OF TECH (JP)
  • US20230376737A1 patent drawing
  • US20230376737A1 patent drawing
  • US20230376737A1 patent drawing

AI summary

A three-dimensional electric element 10 comprises four or more nonlinear units 11 each having nonlinear current-voltage characteristics and an electric conductor 12 connecting the nonlinear units 11, and the nonlinear units 11 are arranged in a three-dimensional manner. A machine learning system 20 comprises a three-dimensional electric element 10 that includes four or more nonlinear units 11 each having nonlinear current-voltage characteristics and an electric conductor 12 connecting the nonlinear units 11 being arranged in a three-dimensional manner, and an input electrode 13 and an output electrode 14, the input electrode 13 and the output electrode 14 are connected to the three-dimensional electric element 10.