Sarcasm-Sensitive Dialog System Using Neural Embeddings

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Spoken dialog systems are ill-equipped to recognize and respond appropriately to sarcastic human communication, often misinterpreting sarcasm and failing to provide relevant responses.

Innovation Solution

A sarcasm-sensitive dialog system that utilizes a neural network to process human speech, incorporating a speech recognition model and a signal knowledge extraction model to detect sarcasm, and generates responses using an embedding vector and one-hot vector to accurately interpret and respond to sarcastic inputs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a traditional spoken dialog system is used, then the system structure is simple, but the system cannot recognize sarcasm and responds inappropriately

Engineering Contradiction:

Improvesarcasm recognition accuracyVSAvoidsystem structure complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The dialog system is segmented into multiple specialized modules: a speech recognition model for transcribing speech, a signal knowledge extraction model for detecting sarcasm, and a response generation model for crafting appropriate replies. Each module focuses on a specific task, improving overall sarcasm recognition accuracy while maintaining manageable complexity through functional decomposition.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

An embedding layer is introduced as an intermediary component that transforms sarcasm detection results into a numerical representation that the response generation model can process. This intermediary bridges the gap between sarcasm detection and response generation, enabling the system to handle sarcastic inputs appropriately without requiring direct complex interaction between detection and generation components.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If sarcasm detection is added to improve response accuracy, then the response relevance improves, but the processing time increases

Engineering Contradiction:

Improveintent recognition precisionVSAvoiddialog processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The speech is transcribed into text and the sarcasm detection is performed on the transcribed text rather than directly on audio signals. This preliminary action of converting speech to text allows for more efficient sarcasm analysis using language processing techniques, reducing the overall processing time while maintaining high intent recognition precision.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system replaces direct audio signal analysis with text-based processing for sarcasm detection. By substituting the mechanical audio processing path with a text-based knowledge extraction path, the system achieves more precise intent recognition while reducing computational complexity and processing time.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11250853B2Sarcasm-sensitive spoken dialog system

Publication Date: 2022.02.15 ROBERT BOSCH GMBH

AI summary

A dialog system and a method of using the dialog system is disclosed. The method may comprise: receiving audible human speech from a user; determining that the audible human speech comprises sarcasm information; providing an input to a neural network, wherein the input comprises speech data input associated with the audible human speech, an embedding vector associated with the sarcasm information, and a one-hot vector; and based on the input, determining an audible response to the human speech.