Acoustic Task Urgency Classification Using Foreground and Background Audio

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing task management systems fail to accurately determine the urgency and importance of tasks based on user interactions, particularly when similar commands are given in different contexts, leading to inconsistent prioritization.

Innovation Solution

A machine learning model analyzes acoustic features of audio data, including foreground and background sounds, to classify tasks into importance and urgency levels using neural networks and regression models, incorporating embeddings and audio sub-streams to enhance accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional task management systems use simple command recognition, then the system is easy to operate, but the accuracy of determining task urgency and importance is poor

Engineering Contradiction:
Improveaccuracy of determining task urgencyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The audio data is segmented into multiple audio sub-streams based on different audio types (e.g., foreground voice, background noise, ambient sounds). Each sub-stream is processed independently to extract relevant acoustic features, allowing the system to analyze different audio components separately and combine their contributions to determine task urgency and importance accurately.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system transitions from analyzing only the textual content of voice commands to analyzing acoustic features across multiple dimensions including audio type, volume, pitch, and temporal characteristics. This dimensional expansion enables the system to capture contextual information about the user's state and environment, significantly improving task prioritization accuracy.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If the system analyzes only voice commands without background context, then the processing is fast, but the determination of task context accuracy is poor

Engineering Contradiction:
Improveaccuracy of task context determinationVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary classification of audio data into different audio types and extracts acoustic features before the main task prioritization process. This preprocessing step organizes the audio data into structured sub-streams, enabling faster and more accurate analysis during task execution without requiring reprocessing of raw audio data.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system extracts specific acoustic features (volume, pitch, tempo, spectral characteristics) from the audio data and separates them from the raw audio signal. By extracting only the relevant features needed for context determination, the system achieves accurate task context analysis while minimizing processing time and computational resources.

Inventive Principle:
Principle #2Taking out (Extraction)

3Measurement precision

If the system uses comprehensive acoustic feature analysis, then the task prioritization accuracy is improved, but the computational complexity increases

Engineering Contradiction:
Improvetask prioritization accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The computational task is divided into separate processing pipelines for different audio types and acoustic features. Each pipeline processes specific audio sub-streams independently, allowing for optimized computation and parallel processing. This segmentation reduces the overall computational complexity while maintaining comprehensive feature analysis.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system implements a hierarchical analysis approach where not all acoustic features are processed with equal depth for every audio input. Based on the audio type and initial feature analysis, the system selectively applies more computationally intensive processing only when necessary, achieving high accuracy while managing computational resources efficiently.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4285362B1Determination of task urgency based on acoustic features of audio data
Publication Date: 2025.08.27 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4285362B1 patent drawingFigure 1
  • EP4285362B1 patent drawingFigure 2
  • EP4285362B1 patent drawingFigure 3

AI summary

Systems and methods are provided for determining importance and urgency of a task based on acoustic features of audio input associated with the task. The determining includes classifying the task into one or more classes associated with importance, urgency, and priority of the task. The classification may use a trained machine learning model of acoustic features and embedding for a neural network. The task classifier uses feature acoustics of either or both the foreground and background audio. The feature acoustics include a pitch, a tone, and a volume over a time duration of the audio input. A combination of the acoustic features determines a class associated with the task. The machine learning model includes a regression model of acoustic features over time and a model with embedding for a neural network.