Entity Grammar for Conversational Agent Dialog Coherence

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current approaches for programming human-computer dialog systems lack coherence and flexibility, making it difficult to create robust and powerful conversational agents that can adapt to user interactions.

Innovation Solution

A method and system using entity grammars that define compound entities in terms of tokens, choice, and product, allowing conversational agents to interpret user utterances and generate prompts, while building a dialog outcome that can adapt and maintain coherence and flexibility through a dialog interpreter.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If finite-state approach or frame-based approach is used for specifying human-computer dialogs, then dialog control can be determined, but coherence and flexibility are not provided

Engineering Contradiction:

Improvecoherence and flexibilityVSAvoiddialog control determination

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The dialog system is segmented into multiple independent components: entity grammars for defining domain concepts, dialog interpreter for managing conversation flow, and outcome builders for constructing results. This segmentation allows each component to specialize, improving both coherence within domains and flexibility across different dialog scenarios while maintaining reliable control through their coordinated interaction.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The entity grammar framework serves multiple functions simultaneously: it defines domain vocabulary, establishes dialog flow rules, and structures outcome data. This multi-functionality eliminates the need for separate finite-state machines and frame definitions, providing coherence through unified grammar rules while enabling flexibility through the same framework's ability to adapt to different dialog types.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Ease of operation

If traditional human-computer interfaces are used, then computers can perform tasks, but interfaces are cumbersome for real-life situations

Engineering Contradiction:

Improveinterface usabilityVSAvoidinterface structure

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The patent replaces traditional mechanical graphical user interfaces with a natural language processing system. Users interact through spoken or written language rather than clicking through complex menu structures. The entity grammar and dialog interpreter automatically parse and respond to natural language, substituting the mechanical interface paradigm with a linguistic one that is inherently more intuitive for real-life situations.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The dialog system enables users to directly express their needs in natural language without navigating complex interface hierarchies. The system's entity grammars and interpreters automatically understand and process these expressions, allowing users to service their own information needs through conversational interaction rather than requiring the system to present structured interface options.

Inventive Principle:

Principle #25Self-service

3Adaptability or versatility

If natural language processing advances are applied, then conversational interfaces can be brought to mainstream, but authoring facilities are lagging behind

Engineering Contradiction:

Improveconversational interface capabilityVSAvoidauthoring difficulty

Core Design Contradiction:

Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

The patent merges the previously separate tasks of dialog flow design, vocabulary definition, and outcome structuring into a unified entity grammar framework. Authors create single grammar files that simultaneously define all aspects of the conversational interface, eliminating the need to separately author finite-state machines, frame definitions, and data structures. This merging dramatically simplifies the authoring process while maintaining advanced NLP capabilities.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The entity grammar framework provides a universal authoring approach that handles multiple functions through a single methodology. The same grammar rules that define domain entities also control dialog flow and structure outcomes, eliminating the need for separate authoring facilities for each function. This universal framework makes creating conversational interfaces accessible to non-experts while supporting sophisticated NLP operations.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS11556704B2Conversational computer agent and outcome

Publication Date: 2023.01.17 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

An entity grammar that specifies a computer conversational agent may be received. User utterances are interpreted based on the entity grammar and prompts for the conversational agent to pose are determined based on the entity grammar. An outcome of the dialog is built by storing words in the user utterances and the prompts that match tokens in the entity grammar. The entity grammar specifies both a dialog flow and data structure of the outcome.