Hybrid Text Image Instruction Generation for Manufacturing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The existing methods for generating textual instructions for manufacturers from designer files are laborious and inaccurate due to differing terminology usage among designers, leading to frequent manual intervention and correction in the manufacturing process.

Innovation Solution

A system and method that utilize a language model to generate textual instructions from hybrid textual and image data, including extracting words from annotated files, determining interrogator outputs from geometric models, and producing accurate manufacturing instructions by simulating and improving upon expert selection processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional manual methods are used to generate textual instructions from designer files, then flexibility in handling diverse designer terminology is maintained, but the process becomes laborious and inaccurate requiring frequent manual intervention

Engineering Contradiction:

Improveaccuracy of manufacturing instructionsVSAvoidautomation of instruction generation

Core Design Contradiction:

ReliabilityVSExtent of automation

Solution Approach 1:

The patent introduces an intermediary system comprising a language model and a geometric model that acts as a mediator between designer files and manufacturing instructions. The language model processes textual data while the geometric model processes image data, and their combined output generates accurate manufacturing instructions. This intermediary system resolves the contradiction by automating the translation process while maintaining accuracy through multiple processing pathways.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent employs a composite approach by combining two different types of models - a language model for textual data and a geometric model for image data. This composite system processes both textual and visual information from designer files simultaneously, leveraging the strengths of each model type to generate more accurate and reliable manufacturing instructions than either model could achieve alone.

Inventive Principle:

Principle #40Composite materials

2Productivity

If automated processing is implemented to reduce manual labor, then productivity increases, but accuracy decreases due to differing terminology usage among designers

Engineering Contradiction:

Improveefficiency of instruction generationVSAvoidaccuracy of manufactured items

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The intermediary processing system with dual models acts as a bridge that maintains precision during automation. The language model handles textual terminology variations while the geometric model validates against visual specifications, ensuring that automated processing does not compromise manufacturing accuracy despite diverse designer terminology.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces manual mechanical processing with an intelligent automated system that uses machine learning models. Instead of manual review and interpretation of designer files, the system automatically processes both text and images through trained models, achieving high productivity while maintaining or improving precision through consistent application of learned patterns from training data.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Manufacturing precision

If expert intervention is used to ensure accurate interpretation of designer files, then manufacturing precision is maintained, but the process becomes laborious and time-consuming

Engineering Contradiction:

Improveaccuracy of manufacturing instructionsVSAvoidtime for instruction generation

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs preliminary action by pre-training both the language model and geometric model on extensive datasets of designer files and corresponding manufacturing instructions. This preliminary training enables the models to automatically interpret diverse terminology and generate accurate instructions without requiring time-consuming expert intervention during actual manufacturing processes.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent substitutes expert human intervention with an automated intelligent system that has been trained to perform the interpretive functions previously requiring expert knowledge. The trained models rapidly process designer files and generate accurate manufacturing instructions, eliminating the time loss associated with manual expert review while maintaining the precision that experts previously provided.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11593563B2Systems and methods for generating textual instructions for manufacturers from hybrid textual and image data

Publication Date: 2023.02.28 PAPERLESS PARTS INC

AI summary

A system for generating textual instructions for manufacturers from hybrid textual and image data includes a manufacturing instruction generator that may generate a language processing module from a first training set including at least a training annotated file describing at least a first product to manufacture, the at least an annotated file containing one or more textual data, and at least an instruction set containing one or more manufacturing instructions to manufacture the at least a first product. Manufacturing instruction generator may use the language processing to generate textual instructions for manufacturers from at least an annotated file and may initiate manufacture using the generated manufacturing instructions.