Information Processing Apparatus for Multi-Intent Question Parsing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing question and answer systems fail to effectively address multiple question intentions within a single query, often providing incomplete or irrelevant answers unless pre-prepared answers are available in the database.

Innovation Solution

An information processing apparatus that parses questions to identify multiple intentions, generates individual questions expressing each intention, and retrieves answers from a database to present relevant responses for each intention, allowing for a more comprehensive answer presentation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a single question is processed to retrieve one answer from the database, then the answer retrieval process is simple and fast, but the system cannot effectively address multiple question intentions within a single query

Engineering Contradiction:
Improveability to handle multiple question intentionsVSAvoidquestion parsing and intention specification complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system segments a single complex question into multiple distinct question intentions through parsing. The question analysis unit divides the input question into several intention components, each representing a separate information need. This allows the system to handle multiple intentions independently while maintaining the ability to provide comprehensive answers for each segment.

Inventive Principle:
Principle #1Segmentation

2Reliability

If the system waits for pre-prepared answers to be available in the database, then answer accuracy is high, but the system cannot provide timely responses when such answers are not available

Engineering Contradiction:
Improveanswer accuracyVSAvoidresponse time delay
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary parsing and intention specification on incoming questions before attempting to retrieve answers. By pre-processing the question into structured intention components, the system prepares the query in advance, enabling faster and more accurate retrieval when answers are available, while also facilitating efficient handling of cases where answers need to be generated or adapted.

Inventive Principle:
Principle #10Preliminary action

3Ease of operation

If the system provides a single comprehensive answer for a multi-intentioned question, then the response structure is simple, but the answer may be incomplete or irrelevant for specific sub-intentions

Engineering Contradiction:
Improveanswer presentation simplicityVSAvoidinformation completeness for individual intentions
Core Design Contradiction:
Ease of operationVSLoss of information

Solution Approach 1:

The system segments the answer presentation to correspond with segmented question intentions. Each parsed intention receives its own dedicated answer or answer component, ensuring that specific information needs are met completely and accurately. This segmented approach maintains clarity and relevance for each sub-intention while collectively addressing the full scope of the original multi-intentioned question.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS20240005101A1Information processing apparatus, information processing method, and non-transitory computer readable medium
Publication Date: 2024.01.04 FUJIFILM BUSINESS INNOVATION CORP
  • US20240005101A1 patent drawing
  • US20240005101A1 patent drawing
  • US20240005101A1 patent drawing

AI summary

An information processing apparatus includes a processor configured to: parse a question to thereby specify a number of intentions included in the question and specify, from the question, one or more question intentions equal to the number; and present an answer retrieved for an individual question expressing each question intention, each answer being presented in association with the corresponding question intention.