Speech Recognition Using Activity Patterns

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current computing devices face challenges in understanding user intent and anticipating user needs due to difficulties in interpreting user speech and recognizing patterns of user activity, leading to frustrated interactions.

Innovation Solution

The system employs sensors on user devices to gather data on user activity patterns over time, analyzing these patterns to predict future actions and infer user intent, which is then used to improve speech recognition and provide personalized experiences by integrating this information into automatic speech recognition and language modeling components.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional speech recognition systems are used without activity pattern analysis, then the system complexity remains low, but the accuracy of understanding user intent deteriorates

Engineering Contradiction:
Improveaccuracy of understanding user intentVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines traditional speech recognition systems with activity pattern analysis systems into a unified framework. The speech recognition component processes user speech while the activity pattern component analyzes sensor data and user behaviors, and both components work together to infer user intent with higher accuracy than either component alone.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces activity patterns as an intermediary element that mediates between raw sensor data and speech recognition output. The activity patterns serve as contextual information that helps disambiguate speech recognition results and improve overall intent understanding without requiring direct complex processing of all sensor data.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If activity pattern analysis is integrated into speech recognition, then the understanding of user speech improves, but the difficulty of detecting and measuring user activity increases

Engineering Contradiction:
Improveaccuracy of speech recognitionVSAvoiddifficulty of detecting user activity
Core Design Contradiction:
Measurement precisionVSDifficulty of detecting and measuring

Solution Approach 1:

The patent segments the activity detection process into distinct components: sensor data collection, activity pattern extraction, and speech recognition enhancement. Each component handles specific tasks independently, making the overall system more manageable and reducing the difficulty of detecting and measuring user activity by breaking it down into smaller, specialized sub-tasks.

Inventive Principle:
Principle #1Segmentation

3Measurement precision

If user activity data is collected and analyzed over time, then the prediction of future user actions improves, but the loss of time for data processing increases

Engineering Contradiction:
Improveaccuracy of predicting future actionsVSAvoidtime for data processing
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary processing of sensor data to extract activity patterns in advance, before speech recognition is needed. By pre-processing and storing activity patterns, the system reduces the time required for data processing during actual speech recognition interactions, as the pattern matching and analysis have already been performed.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10769189B2Computer speech recognition and semantic understanding from activity patterns
Publication Date: 2020.09.08 MICROSOFT TECHNOLOGY LICENSING LLC
  • US10769189B2 patent drawing
  • US10769189B2 patent drawing
  • US10769189B2 patent drawing

AI summary

A user activity pattern may be ascertained using signal data from a set of computing devices. The activity pattern may be used to infer user intent with regards to a user interaction with a computing device or to predict a likely future action by the user. In one implementation, a set of computing devices is monitored to detect user activities using sensors associated with the computing devices. Activity features associated with the detected user activities are determined and used to identify an activity pattern based on a plurality of user activities having similar features. Examples of user activity patterns may include patterns based on time, location, content, or other context. The inferred user intent or predicted future actions may be used to facilitate understanding user speech or determining a semantic understanding of the user.