Intent Recognition via Key Information Extraction and User Confirmation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing intent networks face inefficiencies due to inaccurate user input of service requirements, leading to misaligned network configuration strategies and resource wastage, particularly when users lack experience or understanding of network environments.

Innovation Solution

A method for intent recognition that involves extracting key information from user input, generating a target instance, and allowing for selection, confirmation, and modification to align intent input with actual service requirements, using an electronic device with specific hardware and software modules.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If users input service requirements directly into the intent network, then the network can process requests, but the accuracy of intent recognition deteriorates due to inaccurate or insufficient user input

Engineering Contradiction:
Improveintent recognition accuracyVSAvoiduser input difficulty
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent introduces an intermediary system that includes intent recognition module and network element identifier module. This intermediary extracts key information from user input and generates standardized network element identifiers, serving as a bridge between user-friendly input and precise network configuration, thereby improving intent recognition accuracy without complicating user interaction

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary action by pre-defining multiple network element identifiers and their corresponding service characteristics before user input. The intent recognition module extracts key information and matches it against these pre-defined identifiers, enabling accurate intent recognition even when user input is vague or incomplete

Inventive Principle:
Principle #10Preliminary action

2Productivity

If the intent network processes service requirements without verification, then processing speed is maintained, but resource wastage increases due to misaligned network configuration strategies

Engineering Contradiction:
Improveservice processing efficiencyVSAvoidresource wastage
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent implements feedback mechanisms where the intent recognition module provides feedback on the accuracy of extracted key information, and the network element identifier module verifies whether the generated identifiers align with actual service requirements. This feedback loop enables the system to correct misaligned configurations before execution, reducing resource wastage while maintaining processing efficiency

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system replaces direct mechanical processing of user input with an automated information extraction and matching mechanism. The intent recognition module uses information extraction techniques to identify key parameters, and the network element identifier module uses pattern matching to select appropriate identifiers, substituting manual verification processes with automated intelligent processing that reduces errors and resource waste

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Ease of operation

If users lack experience or understanding of network environments, then ease of operation is improved for users, but intent recognition accuracy deteriorates due to insufficient or incorrect input

Engineering Contradiction:
Improveuser input simplicityVSAvoidservice requirement accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent enables the system to perform self-service by automatically extracting key information from user input and generating appropriate network element identifiers without requiring user expertise. The intent recognition module autonomously identifies relevant parameters, and the network element identifier module autonomously selects matching identifiers, allowing inexperienced users to provide simple input while the system handles the complex processing independently

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system applies parameter changes by transforming user input parameters into standardized network element identifier parameters. The intent recognition module extracts key information in a standardized format, and the network element identifier module transforms this into corresponding identifier parameters, enabling the system to handle various user input formats and convert them into precise network configuration parameters

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4697665A1Intent recognition method, device, and storage medium
Publication Date: 2026.02.18 ZTE CORP
  • EP4697665A1 patent drawingFigure 1~2
  • EP4697665A1 patent drawingFigure 3~4
  • EP4697665A1 patent drawingFigure 5~6

AI summary

Provided are an intent recognition method, a device, and a storage medium. The method comprises: receiving target information, wherein the target information comprises a service demand of a target user; extracting key information in the target information, wherein the key information comprises: service implementation position information, service implementation time information, service feature information, an identifier of a logical network element for implementing a service, an identifier of a network device for implementing a service, an identifier of a terminal device for implementing a service, and/or an identifier of application software for implementing a service; on the basis of the key information, generating and outputting a target instance, wherein the target instance is used for representing intent input information; and receiving an operation on the target instance to complete intent recognition, wherein the operation comprises selection, confirmation, and/or modification.