Dialogue System Keyword Weighting for Intent Accuracy

Solution Overview

Problem

Existing dialogue systems struggle to accurately interpret user intentions due to erroneous input keywords, leading to inefficiencies in narrowing down questions.

Innovation Solution

A dialogue system that prompts users for additional input, determines priority rankings based on the degree of coincidence between input keywords and stored keywords, and reduces the influence of keywords with additional information to correctly understand user intentions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If keyword correction is applied to handle erroneous input, then the system can accommodate user input variations, but it may take time to narrow down questions when correction is not in line with user intention

Engineering Contradiction:

Improvehandling of erroneous inputVSAvoidtime to narrow down questions

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system applies keyword correction partially by reducing the influence weight of corrected keywords rather than completely replacing them. This allows the system to benefit from correction when appropriate while preserving original keywords that may still be relevant, thus avoiding excessive time consumption in narrowing down questions.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system changes the parameter of keyword influence by assigning different weights to keywords based on their correction status. Corrected keywords are given reduced influence weights, while uncorrected keywords maintain full weight, allowing dynamic adjustment of keyword importance based on input reliability.

Inventive Principle:

Principle #35Parameter changes

2Reliability

If additional information is added to input keywords to handle notation fluctuations, then erroneous inputs can be corrected, but the degree of coincidence calculation may be affected negatively

Engineering Contradiction:

Improvecorrection of notation fluctuationVSAvoiddegree of coincidence

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The system changes the parameter of keyword influence by assigning different weights to keywords based on their correction status. Corrected keywords are given reduced influence weights, while uncorrected keywords maintain full weight, allowing dynamic adjustment of keyword importance based on input reliability.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system introduces an intermediary mechanism (influence weight) between keyword matching and final ranking. This intermediary allows corrected keywords to contribute to the matching process without fully determining the outcome, thus maintaining measurement precision while handling notation fluctuations.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Productivity

If the system selects resulting sentence candidates based on degree of coincidence, then relevant questions can be identified, but intended questions may be buried when erroneous keywords are present

Engineering Contradiction:

Improvequestion identification efficiencyVSAvoidaccuracy of intention understanding

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system changes the parameter of keyword influence by assigning different weights to keywords based on their correction status. Corrected keywords are given reduced influence weights, while uncorrected keywords maintain full weight, allowing dynamic adjustment of keyword importance based on input reliability.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system uses feedback from the ranking determination process to adjust the influence of keywords. By evaluating the degree of coincidence with reduced influence weights for corrected keywords, the system can identify whether the correction led to appropriate matching or if original keywords should have been preserved, effectively using the outcome to validate the correction process.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11663420B2Dialogue system

Publication Date: 2023.05.30 NTT DOCOMO INC

AI summary

The dialogue system includes a ranking determination unit configured to determine a priority ranking of the resulting sentence candidates on the basis of a degree of coincidence between the input key group and the storage key group, and a presentation unit configured to present a response for prompting a user to make an additional input on the basis of a resulting sentence candidate of a highest priority ranking. In a case where there is the stored keyword which is in a predetermined correspondence relation with the character string, the keyword acquisition unit acquires the input keyword obtained by adding additional information to the stored keyword. In a case where the additional information is added to the input keyword, the ranking determination unit reduces an influence of the input keyword to which the additional information is added on the degree of coincidence.