Agenda-Driven Question Generation for Goal-Aligned Dialogs

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing natural language processing systems lack the ability to generate questions that effectively further the goals of a dialog, particularly in contexts such as legal examinations, dispute resolutions, and professional interviews, by leveraging specific corpora of information to enhance user interactions.

Innovation Solution

A system is configured to generate and output questions that further a dialog's goal using a corpus of information, incorporating components like an agenda-driven question generation component and a trained machine learning model to process user inputs and generate relevant questions based on dialog history and knowledge data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If existing natural language processing systems are used, then basic speech recognition and natural language understanding are achieved, but the ability to generate questions that effectively further dialog goals is lacking

Engineering Contradiction:

Improveability to generate questions aligned with dialog goalsVSAvoidsystem complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system divides the question generation task into separate functional modules: speech recognition component, natural language understanding component, and natural language generation component. Each module processes specific aspects of the dialog, allowing the system to generate goal-aligned questions without requiring complete redesign of the entire NLP pipeline.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary processing of user inputs through speech recognition and natural language understanding before generating questions. By analyzing the dialog history and identifying the current goal in advance, the system can generate relevant questions that advance the dialog objectives rather than responding reactively to each input.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If a trained machine learning model is used to generate questions, then question relevance to dialog goals improves, but processing time and computational resources increase

Engineering Contradiction:

Improvequestion relevance to dialog goalsVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system applies natural language generation only when needed to generate specific questions based on identified dialog goals, rather than continuously processing all input data through the full ML pipeline. This selective application of the trained model reduces unnecessary computational overhead while maintaining high question relevance when question generation is required.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12456015B1Natural language question generation

Publication Date: 2025.10.28 AMAZON TECH INC

AI summary

Techniques for generating a natural language prompt to further a goal of a dialog, are described. During a dialog, the system receives one or more user inputs including a user question, a user response to the question, and a request to generate a further question following the response. The system determines ASR output data corresponding to the user inputs, and determines dialog history data of the dialog. Using the ASR output data and the dialog history data, the system determines a category and an explanation of relevance corresponding to the category. Using the ASR output data, the dialog history, the category, and the explanation, the system determines the further question to be output to the user.