Specification Preparation Workflow With Chapter Revision Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for preparing patent specifications lack consideration for high-quality specifications, ease of patent grant, ease of exercising rights, and time efficiency, particularly in systems utilizing language models like Retrieval-Augmented Generation (RAG).
Innovation Solution
A system for supporting specification preparation that includes components to receive a summary of an invention, generate prompts, classify and revise chapters based on a language model, and compare chapters to select suitable content for inclusion in the specification, utilizing a database for data retrieval and revision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a language model is used to generate patent specifications automatically, then preparation time is reduced, but the quality and completeness of the specification may be insufficient
Solution Approach 1:
The system segments the specification generation process into multiple iterative steps: initial generation, chapter classification, data retrieval, revision, and comparison. Each step focuses on specific aspects of specification quality, allowing the language model to produce high-quality specifications while maintaining efficient automation through structured processing stages.
Solution Approach 2:
The system implements feedback mechanisms where generated specifications are classified into chapters, compared against retrieved data, and revised iteratively. The comparison step provides feedback on completeness and accuracy, guiding subsequent revisions to improve specification quality while maintaining automated preparation.
2Productivity
If existing systems generate specifications without considering patent grant criteria, then generation is simpler and faster, but the ease of patent grant and scope of claims are reduced
Solution Approach 1:
The system performs preliminary actions by retrieving relevant patent data and examination criteria before finalizing the specification. Chapters are classified and data is obtained in advance, allowing the specification to be drafted with consideration for patent grant requirements, ease of exercising rights, and claim scope from the outset rather than as an afterthought.
Solution Approach 2:
The system changes parameters by adjusting specification content based on retrieved data and classification results. The revision process modifies technical descriptions, claim scope, and supporting details to optimize for patent grant probability and claim breadth, transforming the specification from a basic technical description to a strategically optimized patent document.
3Manufacturing precision
If detailed data retrieval and chapter revision are performed, then specification quality improves, but the time and complexity of preparation increase
Solution Approach 1:
The system divides the complex specification preparation process into manageable segments: initial generation, chapter classification, data retrieval, revision, and comparison. This segmentation reduces perceived complexity by organizing tasks into clear, sequential steps while maintaining comprehensive quality control through each stage.
Solution Approach 2:
The system employs multi-functional components that perform multiple operations. The language model generates, revises, and compares specifications; the classification system organizes content and retrieves data; and the integration layer coordinates all functions. This multi-functionality reduces overall system complexity by consolidating operations rather than requiring separate specialized systems for each task.
Data Source
AI summary
A system for supporting specification preparation having a function of receiving a summary of an invention, generating a first prompt including a content for preparing a first specification corresponding to the summary, obtaining the first specification from a language model as an answer to the first prompt, displaying the first specification, classifying the first specification into chapters, generating a second prompt including a content for obtaining data related to a first chapter that is classified from a second component and a content for creating a second chapter by revising the first chapter in accordance with the data, obtaining the second chapter as an answer to the second prompt from the language model, displaying the first chapter and the second chapter, comparing the first chapter and the second chapter to select the one which is suitable for the summary, and preparing a second specification including the selected chapter is provided.