Contextual Meeting Transcripts With ML Tone and Direction Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for inferring the context of meetings, such as audio/visual recordings and manual transcription, are resource-intensive and inefficient, and may raise privacy concerns, while human interpretation of transcripts can lead to inaccurate assumptions about tone and context.
Innovation Solution
A computer system using machine learning models trained on user utterances to automatically detect tone and direction in meeting transcripts, integrating tone and direction data directly into the transcript for improved context understanding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If audio/visual recordings are made to capture meeting context, then context information is preserved, but resource consumption and storage requirements become prohibitive
Solution Approach 1:
The patent extracts only the essential context information (tone and direction) from the meeting data, separating it from the full audio/visual recording. Machine learning models process the complete meeting data to extract these specific contextual features, which are then stored separately from the original recordings, reducing storage requirements while preserving essential context.
Solution Approach 2:
Instead of storing original audio/visual recordings, the system creates simplified copies in the form of structured context data (tone labels, direction indicators) that capture the essential meeting dynamics. These contextual copies are stored in a compact format that requires minimal storage resources while maintaining the ability to reconstruct meeting context.
2Reliability
If manual review of transcripts is performed to add context information, then context is captured, but time consumption increases significantly
Solution Approach 1:
The patent replaces the manual mechanical process of human review with automated machine learning models. These models process transcripts and detect tone and direction context automatically, eliminating the need for human reviewers to manually analyze each meeting transcript while maintaining or improving context detection accuracy through consistent application of trained algorithms.
Solution Approach 2:
The system enables transcripts to self-analyze through machine learning models that automatically process the text and extract context information. The models are trained to independently identify tone and direction without human intervention, allowing the system to serve itself in generating contextualized transcripts at scale.
3Loss of information
If human reviewers add context to transcripts, then context information is provided, but accuracy of tone interpretation may be false
Solution Approach 1:
The patent transforms subjective human interpretation into objective measurable parameters through machine learning models. The models convert tone and direction into standardized categorical labels based on linguistic patterns and voice characteristics, removing the subjectivity and variability inherent in human interpretation while providing consistent, measurable context information.
Solution Approach 2:
The system incorporates feedback mechanisms where machine learning models are trained on labeled data and continuously improve their tone and direction detection accuracy. The models learn from correct and incorrect interpretations, adjusting their parameters to reduce false assumptions about speaker tone while maintaining high accuracy in context detection.
Data Source
AI summary
Systems and methods are disclosed herein for building contextual transcripts. A computing system may receive a textual transcript of a meeting that contains a variety of statements made by various attendees of the meeting, select the first statement made during the meeting, and determine which meeting attendee made the statement. A machine learning model corresponding to the particular attendee that has been trained using previously received statements by the particular attendee may be used on the utterance to determine the tone of the utterance. That tone may be recorded within the transcript and this process may be repeated for each utterance to build a contextual transcript.