Automated Patent Drafting System Using Vectorized Seed Sentences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The conventional process of drafting patent applications is laborious and expensive, requiring significant manual effort to synthesize invention disclosures into specifications and claims, despite efforts to increase efficiency through forms and templates.
Innovation Solution
A document generation system that uses a model generator and document generator to create patent applications based on seed sentences, employing vectorization and sequence generation models to automate the drafting process, reducing manual labor and increasing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual synthesis of invention disclosures into patent specifications and claims is performed, then the quality and accuracy of the patent application can be maintained, but the labor time and cost increase significantly
Solution Approach 1:
The patent application drafting system acts as an intermediary between the inventor's disclosure and the final patent application. It automatically processes invention disclosures, extracts technical features, and generates patent specifications and claims, reducing manual intervention while maintaining quality through structured processing and expert rule-based generation
Solution Approach 2:
The system enables self-service patent drafting by allowing inventors to input their disclosures through standardized forms, and the system automatically generates the patent application documents without requiring manual synthesis by patent attorneys or drafters, thereby reducing labor time while maintaining acceptable quality through automated processing
2Ease of operation
If invention disclosure forms with abstractions and roadmaps are provided to inventors, then the organization of information improves, but significant manual effort is still required to author the specification and claims
Solution Approach 1:
The system performs preliminary actions by pre-processing the invention disclosure information through automated extraction and analysis before patent document generation. It identifies technical features, relationships, and key elements from the disclosure forms, preparing structured data that can be directly used for automated specification and claim generation, thereby eliminating the need for manual information reorganization
Solution Approach 2:
The system replaces the mechanical manual process of synthesizing disclosure information into patent documents with an automated computational system. It uses natural language processing, pattern recognition, and rule-based generation to convert disclosure forms into patent specifications and claims, substituting human manual effort with automated intelligence while maintaining productivity
3Ease of manufacture
If patent templates with pre-written sections are used, then the drafting process is facilitated, but significant manual effort remains to author the specification and claims
Solution Approach 1:
The system uses copying by extracting and reusing standardized patent language, claim structures, and specification templates from existing patent databases. It automatically copies proven patterns and formulations into the generated patent application, eliminating the need for manual authoring while maintaining the quality and consistency associated with established patent drafting conventions
Solution Approach 2:
The system achieves universality by creating a multi-functional patent drafting platform that can handle various types of invention disclosures, generate different sections of patent applications (specification, claims, abstracts), and adapt to different technical fields. This single system replaces multiple manual processes and template-based approaches, significantly reducing manual effort while maintaining ease of use through standardized interfaces
Data Source
AI summary
A document generation system to generate a document that forms the basis for a patent application to be submitted for examination by a patent searching authority. The document generation system may include: 1) a model generator trained on a plurality of references obtained from a reference database and 2) a document generator configured to generate paragraphs of a document based on a trained model.