Context-Aware User Action Prediction via Semantic Vector Indexing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional mobile communication devices lack context awareness and predictive capabilities, presenting users with static and irrelevant options, failing to adapt to their current situation or context.

Innovation Solution

A processor-based system that detects user actions, translates them into semantic vectors, and uses a random index algorithm to predict next actions by comparing current context information with past data, allowing for dynamic and context-aware predictions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional mobile devices present pre-defined static choices to users, then the device operation is simple and straightforward, but the device lacks context awareness and cannot adapt to user situations

Engineering Contradiction:
Improvecontext awarenessVSAvoidprediction system complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by continuously monitoring user actions, time, and location data in the background, building a knowledge base of user behavior patterns before predictions are needed. This allows the device to proactively prepare prediction models rather than reacting only when predictions are requested.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary prediction system that acts as a mediator between raw sensor data (user actions, time, location) and the user interface. This intermediary layer processes and interprets data to generate contextual predictions, shielding the complexity of the prediction algorithm from both the hardware and the user.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Loss of information

If all possible choices are presented to the user, then the user has complete information, but the user is overwhelmed with irrelevant information in certain situations

Engineering Contradiction:
Improveinformation completenessVSAvoiduser interface simplicity
Core Design Contradiction:
Loss of informationVSEase of operation

Solution Approach 1:

The system applies local quality by customizing the information presentation based on the specific context (location, time, user action). Instead of uniformly presenting all possible choices everywhere, the interface adapts locally to show only relevant options for the current situation, making the information both complete and contextually appropriate.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent implements partial action by selectively presenting only a subset of all possible choices - specifically those most relevant to the current context. Rather than showing all available options (excessive), the system shows just enough information (partial) to help the user make informed decisions without overwhelming them.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If the device monitors multiple parameters (user action, time, location) to improve prediction accuracy, then prediction quality improves, but the processing complexity and energy consumption increase

Engineering Contradiction:
Improveprediction accuracyVSAvoidprocessing energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system applies partial action by monitoring all relevant parameters (user action, time, location) but using them selectively based on context. Not all parameters are processed with equal depth at all times - the system adjusts the level of analysis based on what is currently needed for accurate prediction, reducing unnecessary processing energy.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent implements parameter changes by dynamically adjusting the monitoring and processing intensity of different parameters based on the situation. For example, location data may be monitored continuously for navigation contexts but less frequently for other activities, optimizing the balance between prediction accuracy and energy consumption.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3107317B1User activity random indexing predictor for mobile context awareness
Publication Date: 2018.11.21 HUAWEI TECH CO LTD
  • EP3107317B1 patent drawingFigure 1
  • EP3107317B1 patent drawingFigure 2
  • EP3107317B1 patent drawingFigure 3

AI summary

An apparatus for predicting a next user action includes a processor configured to detect a user action, detect at least one of a time and a location of the user action, translate the user action and at least one of the time of the user action and the location of the user action into a set of words, calculate a semantic similarity between the set of words and at least one previously generated set of words and determine one or more predicted actions of the user based on a closest semantic similarity between the set of words and the at least one previously generated set of words.