Spoken Dialogue System Query Analysis via Modular Apparatus Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current spoken dialogue systems face challenges in understanding natural language at a near-human level, particularly in processing queries that may contain global commands, out-of-domain queries, and determining intent and domain probabilities, which affects their ability to perform accurate actions and confirmations.

Innovation Solution

A spoken dialogue system is designed with multiple apparatuses to analyze queries for global commands, out-of-domain queries, intent and domain probabilities, and entity tagging, using word and phrase embeddings to improve understanding and accuracy, and includes a confirmation apparatus to verify user actions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a spoken dialogue system uses multiple apparatuses to analyze queries for global commands, out-of-domain queries, intent and domain probabilities, and entity tagging, then the understanding accuracy of natural language is improved, but the device complexity increases

Engineering Contradiction:
Improveunderstanding accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The spoken dialogue system is divided into multiple specialized apparatuses: a global command determination apparatus, an out-of-domain query determination apparatus, an intent and domain determination apparatus, an entity tagging apparatus, and a confirmation apparatus. Each apparatus handles a specific aspect of query analysis, allowing the system to improve understanding accuracy through specialized processing while managing complexity through modular design.

Inventive Principle:
Principle #1Segmentation

2Productivity

If the spoken dialogue system processes global commands, out-of-domain queries, and intent determination simultaneously, then the productivity of query processing is improved, but the device complexity increases

Engineering Contradiction:
Improvequery processing efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system performs preliminary classification by determining whether a query is a global command or an out-of-domain query before proceeding to intent and domain analysis. This preliminary action filters out queries that do not require full processing, improving overall query processing efficiency while reducing unnecessary computational complexity for simple queries.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts its processing depth based on the query type. For global commands, it performs only global command determination; for out-of-domain queries, it stops after that classification; and only for in-domain queries does it proceed to full intent, domain, and entity analysis. This dynamic processing optimizes productivity while managing complexity adaptively.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3652733B1Contextual spoken language understanding in a spoken dialogue system
Publication Date: 2022.04.20 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3652733B1 patent drawingFigure 1
  • EP3652733B1 patent drawingFigure 2
  • EP3652733B1 patent drawingFigure 3

AI summary

A spoken dialogue system includes a spoken language understanding apparatus. The spoken language understanding apparatus can include an intent apparatus and a selection apparatus. The intent apparatus is configured to determine if a query comprises a global command, to determine if an intent associated with a query is or is not included in a domain that is supported by the spoken dialogue system, to determine if a query comprises a confirmation type, to tag one or more entities in a query, and to determine an intent probability distribution and a domain probability distribution that is associated with a query. When the query includes an entity that is included in two or more possible entities, the selection apparatus is configured to provide a score for each of the two or more possible entities.