Generative AI Question Answering System with Provider Attribute Display

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing technologies face challenges in facilitating the use of generative artificial intelligence models, particularly in effectively transferring and utilizing tacit knowledge between users with different skill levels and across various tasks.

Innovation Solution

An information processing apparatus and method that includes storage means for a model generated using text data as training, text information generation means to generate answers based on questions and the model, and screen generation means to display attribute information of data providers, enhancing the usability of the model.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a generative AI model is used to provide answers, then the model's capabilities are utilized, but it is difficult to facilitate the use of the model and transfer tacit knowledge between users

Engineering Contradiction:
Improvemodel usabilityVSAvoidease of use
Core Design Contradiction:
Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent introduces an intermediary system that acts as a bridge between the generative AI model and users. This intermediary extracts and manages tacit knowledge from the model's outputs, organizing it into reusable formats that facilitate knowledge transfer between users while maintaining the model's core functionality

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent extracts tacit knowledge from the generative AI model's responses and separates it as a distinct, manageable component. This extracted knowledge can be independently utilized, shared, and transferred between users, making the model's value more accessible and easier to operate with

Inventive Principle:
Principle #2Taking out (Extraction)

2Reliability

If conventional AI is used to present the best answer from learned data, then accurate answers are provided, but the system lacks the ability to output original content that has not been input

Engineering Contradiction:
Improveanswer accuracyVSAvoidcontent generation capability
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent merges the strengths of conventional AI (accurate retrieval from learned data) with generative AI capabilities (creating original content). The system combines reliable answer presentation with the ability to generate novel content, achieving both accuracy and creativity in a unified system

Inventive Principle:
Principle #5Merging (Combining)

3Adaptability or versatility

If the model continuously learns from various inputs, then original content is generated, but it becomes difficult to facilitate the use and control of the model

Engineering Contradiction:
Improvecontinuous learning capabilityVSAvoidmodel management complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the continuous learning process into manageable components, organizing learned knowledge into structured tacit knowledge units. This segmentation makes the continuously evolving model more controllable and easier to manage while preserving its adaptive learning capabilities

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4553674A1Question answering method and system based on trained model
Publication Date: 2025.05.14 RICOH CO LTD
  • EP4553674A1 patent drawingFigure 1
  • EP4553674A1 patent drawingFigure 2
  • EP4553674A1 patent drawingFigure 3

AI summary

An information processing apparatus includes: storage means (4000) configured to store a model generated using text data as training data, the text data being data based on at least one of voice information and character information; text information generation means (45) configured to generate, based on a question and the model, text information indicating an answer to the question; and screen generation means (42) configured to generate a screen based on attribute information of one or more data providers who have provided the at least one of voice information and character information.