AI Model Generation Assistance via Trial Association Visualization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for constructing prediction models lack clarity in the process, making it difficult to understand and assist in the generation of models with higher accuracy.

Innovation Solution

A model generation assistance apparatus and method that acquires trial information, infers associations between trials based on differences in this information, and outputs display data with nodes representing trials and links indicating associations, facilitating a more understandable visualization of the model construction process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If the model construction process is repeated while changing construction conditions to achieve higher accuracy, then the model accuracy is improved, but the understandability of the construction process deteriorates

Engineering Contradiction:
Improvemodel accuracyVSAvoidunderstandability of construction process
Core Design Contradiction:
Manufacturing precisionVSLoss of information

Solution Approach 1:

The patent creates a visual copy or representation of the model construction process through a diagram that displays multiple trials and their associations. This visual copy preserves the essential information about what processes were taken while simplifying the complex repeated construction activities into an understandable graphical format, resolving the contradiction between maintaining detailed accuracy improvement processes and ensuring process understandability

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent introduces an intermediary visualization system that mediates between the complex model construction process and the user's understanding. The diagram acts as an intermediary representation that translates complex trial information into an easily understandable format, allowing users to grasp the construction process without being overwhelmed by the complexity of repeated model construction activities

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If multiple trials are conducted to construct prediction models with higher accuracy, then the model performance is improved, but the complexity of tracking and understanding different trials increases

Engineering Contradiction:
Improveprediction model performanceVSAvoidcomplexity of tracking trials
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the complex model construction process into discrete trials, each represented as a separate node in the diagram. This segmentation allows multiple trials to be tracked independently while maintaining overall process visibility, reducing the complexity of tracking by breaking down the monolithic construction process into manageable, visually distinct units

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from a one-dimensional list or tabular representation of trials to a two-dimensional graphical diagram where trials are represented as nodes with spatial relationships. This dimensional change enables better visualization of associations between trials, making it easier to track and understand the complex relationships without increasing perceived complexity

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

Data Source

PatentUS20240281679A1Model generation assistance device, model generation assistance method, and program
Publication Date: 2024.08.22 NEC CORP
  • US20240281679A1 patent drawing
  • US20240281679A1 patent drawing
  • US20240281679A1 patent drawing

AI summary

In order to attain the object of providing a technique that allows a process of constructing a model to be presented in an easily understandable manner, a model generation assistance apparatus includes: an acquisition section that acquires trial information including a parameter used in a trial in a process of constructing an AI model; an inference section that infers association between a plurality of trials on the basis of a difference between respective pieces of trial information of the plurality of trials; and an output section that outputs display data including (i) a plurality of nodes respectively representing the plurality of trials and (ii) a link representing the association.