Abstracted Similar Documents for Confidential Text Generation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing document creation assistance devices risk leaking confidential information when generating output text from documents stored in databases.

Innovation Solution

An information processing method that acquires input text, generates an abstracted similar document by abstracting confidential information using a conceptual information tree, and uses a text generation model to produce answer text based on the abstracted document.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If documents with confidential information are stored in the database and used for text generation, then the quality and relevance of output text is improved, but the risk of confidential information leakage increases

Engineering Contradiction:

Improveoutput text qualityVSAvoidconfidential information leakage risk

Core Design Contradiction:

ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent extracts confidential information from documents stored in the database before using them for text generation. By removing sensitive data while retaining the structural and contextual value of the documents, the system maintains output quality while eliminating leakage risks

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces an intermediary processing step that mediates between the original confidential documents and the text generation model. This intermediary layer processes documents to remove sensitive information while preserving useful content, acting as a buffer that protects against information leakage

Inventive Principle:

Principle #24Intermediary (Mediator)

2Object-affected harmful factors

If confidential information is abstracted from documents before processing, then the security against information leakage is improved, but the amount of useful information available for generation is reduced

Engineering Contradiction:

Improveinformation leakage protectionVSAvoiduseful information loss

Core Design Contradiction:

Object-affected harmful factorsVSLoss of information

Solution Approach 1:

The patent segments information into confidential and non-confidential portions, selectively abstracting only the sensitive elements while preserving the rest. This selective segmentation allows the system to protect against leakage while maintaining the usefulness of the document content for text generation

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20250355911A1Information processing method and recording medium

Publication Date: 2025.11.20 SCREEN HOLDINGS CO LTD

AI summary

The invention provides a technique capable of effectively reducing leakage of confidential information that is caused when a text generation model outputs text including the confidential information. The information processing method includes a) acquiring input text, b) acquiring an abstracted similar document based on a document registered in a document database, the abstracted similar document being similar to the input text and including confidential information abstracted by abstraction processing, and c) acquiring output text by using a text generation model, the output text being answer text for the input text when the input text and the abstracted similar document have been input, the text generation model being trained to generate answer text based on text and external information associated with the text.