Architecture Diagram Generation Using Machine Learning Assessment

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

System architects face challenges in creating accurate architecture diagrams, as issues during implementation are often not addressed in the design phase, leading to repeated problems in future system implementations, and lessons learned are not effectively conveyed to future architects.

Innovation Solution

A system comprising an input processor, machine learning processor, and advice generator that assesses entity interconnections using a training dataset to provide recommendations for correcting architecture diagrams, improving accuracy and relevance of feedback over time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If system architects manually create and revise architecture diagrams during implementation, then problems can be corrected in the current project, but lessons learned are not conveyed to future architects creating diagrams for different systems

Engineering Contradiction:

Improveaccuracy of architecture diagramVSAvoidloss of lessons learned

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The system implements feedback by capturing implementation problems and corrections, storing them in a training dataset, and using machine learning to generate recommendations that feed back into future architecture diagram creations. This closes the loop between implementation experiences and design improvements across different projects.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system copies lessons learned from implementation problems and stores them as training data. The machine learning model then copies these patterns to generate recommendations for future architecture diagrams, transferring knowledge across different system implementations without requiring manual replication of lessons.

Inventive Principle:

Principle #26Copying

2Ease of manufacture

If developers revise schematics to correct erroneous portions during implementation, then immediate problems are resolved, but the same problems are likely to be made again if the same schematics are used in new system implementations

Engineering Contradiction:

Improveease of implementing correctionsVSAvoidrepeated implementation of same problems

Core Design Contradiction:

Ease of manufactureVSProductivity

Solution Approach 1:

The system performs preliminary action by analyzing implementation problems and generating corrected architecture diagrams before new implementation projects begin. The machine learning model proactively identifies potential errors and suggests corrections in advance, preventing the repetition of mistakes rather than merely reacting to them after they occur.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system enables self-service by allowing the machine learning model to automatically analyze implementation problems, learn from them, and generate corrective recommendations without requiring manual intervention to transfer lessons between projects. The system serves itself by continuously improving its own recommendation capabilities through the training dataset.

Inventive Principle:

Principle #25Self-service

3Reliability

If multiple versions of architecture diagrams are created during implementation revisions, then problems can be addressed, but hardware demands and memory utilization increase

Engineering Contradiction:

Improvecorrectness of architecture diagramVSAvoidnumber of diagram versions

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The system extracts the essential corrections and lessons from implementation problems, separating the valuable knowledge from the redundant diagram versions. Instead of storing multiple complete diagram versions, the system extracts key corrections into a training dataset that feeds the machine learning model, eliminating the need to maintain multiple full copies of diagrams.

Inventive Principle:

Principle #2Taking out (Extraction)

4Ease of operation

If architects create architecture diagrams without automated assessment, then the creation process is simple, but the accuracy and relevance of the diagrams suffer

Engineering Contradiction:

Improvesimplicity of diagram creationVSAvoidaccuracy of entity interconnections

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system introduces an intermediary machine learning assessment module that sits between the architect's diagram creation and the final diagram output. This intermediary automatically evaluates entity interconnections against the training dataset, providing accuracy enhancement without significantly complicating the creation process, as the assessment occurs automatically in the background.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11018949B2System for generating an architecture diagram

Publication Date: 2021.05.25 ACCENTURE GLOBAL SOLUTIONS LTD

AI summary

A system for generating an architecture diagram includes an input processor, a machine learning processor, and an advice generator. The input processor is configured to receive, from a terminal, entity data associated with a plurality of entities of an architecture and path data associated with a plurality of paths that correspond to interconnections between the plurality of entities. The machine learning processor utilizes a training dataset to assess whether the entities defined by the entity data are correctly interconnected as defined by the path data. The advice generator receives the assessment from the machine learning processor, prepares a recommendation based on the assessment, and communicates the recommendation to the terminal. User feedback is represented in the training data to improve the relevancy of the recommendation.