Deep Learning Network for Objective Voice Trust Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies lack an objective, reproducible, and automated method to quantify and improve trust in conversations, relying heavily on subjective ratings and lacking a standardized approach to measure trust in interpersonal relationships.
Innovation Solution
The use of deep learning models to construct a system that quantifies trust scores by obtaining voice samples, generating predicted trust scores through a deep-learning network with multiple branches and an aggregation network, and training the network based on the predicted and actual trust scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If subjective ratings are used to assess trust, then trust assessment can be performed, but the measurement lacks objectivity and reproducibility
Solution Approach 1:
The patent replaces subjective human rating mechanisms with an automated deep learning system that processes audio signals. The deep learning model analyzes voice samples and generates trust scores objectively, eliminating the need for human subjective assessment while maintaining measurement precision.
Solution Approach 2:
The system uses self-training mechanisms where the deep learning model automatically learns from audio data and generates trust scores without requiring external human intervention for each assessment. The model continuously improves through training on more data, making the system self-sufficient and reproducible.
2Measurement precision
If deep learning models with multiple branches are used to quantify trust, then measurement precision improves, but device complexity increases
Solution Approach 1:
The deep learning network is divided into multiple independent branches, each processing different aspects of audio data (e.g., tone, rhythm, content). This segmentation allows the system to achieve high measurement precision by analyzing multiple dimensions simultaneously while maintaining manageable complexity through modular architecture.
Solution Approach 2:
Multiple branches of the deep learning network are merged through an aggregation layer that combines their outputs into a final trust score. This merging mechanism integrates the strengths of each branch while distributing the computational load, achieving accurate trust quantification without requiring a single overly complex model.
3Productivity
If automated trust scoring is implemented, then productivity increases, but the system requires complex deep learning training processes
Solution Approach 1:
The system performs preliminary training on a comprehensive dataset of audio samples and trust scores before actual use. This pre-training phase establishes the foundation for rapid, accurate trust assessment during operation. The model learns patterns and relationships in advance, enabling efficient automated scoring without requiring time-consuming analysis during deployment.
Solution Approach 2:
The system incorporates feedback mechanisms where predicted trust scores are compared with actual trust assessments, and the model is continuously refined. This feedback loop allows the system to improve its accuracy over time while maintaining high productivity during operation. The feedback mechanism ensures the model adapts to different contexts and improves its training efficiency.
Data Source
AI summary
The present disclosure relates to a system, a method, and a product for using deep learning models to quantify and/or improve trust in conversations. The system includes a non-transitory memory storing instructions executable to construct a deep-learning network to quantify trust scores; and a processor in communication with the non-transitory memory. The processor executes the instructions to cause the system to: obtain a trust score for each voice sample in a plurality of audio samples, generate a predicated trust score by the deep-learning network based on each voice sample in the plurality of audio samples, wherein the deep-learning network comprises a plurality of branches and an aggregation network configured to aggregate results from the plurality of branches, and train the deep-learning network based on the predicated trust score and the trust score for each voice sample to obtain a training result.