Touch-Screen Text Prediction Engine for Mobile Input Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current text input methods for electronic devices, such as QWERTY-style keyboards and handwriting recognition, face limitations in accuracy and user efficiency, especially in noisy environments and for mobile devices, where traditional character-based input systems are not optimized for word or phrase prediction.

Innovation Solution

A touch-screen interface utilizing a text prediction engine that generates predictions concurrently from multiple language models, including human language and application-specific models, using n-gram maps and probabilistic tries to provide adaptive and context-aware word or phrase suggestions, reducing user labor and improving input efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If traditional QWERTY-style keyboards are used for text input, then users can type with familiarity and training, but the input process requires multiple keystrokes per character and is not optimized for mobile devices

Engineering Contradiction:
Improvetyping familiarityVSAvoidinput speed
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The system performs preliminary action by predicting the user's intended word or phrase before the user completes typing. The prediction engine analyzes partial input and proactively generates candidate completions, allowing users to select predicted text rather than type entire words, thus reducing keystrokes while maintaining ease of operation

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The text input system provides self-service through automatic word completion and prediction. The prediction engine continuously monitors user input and automatically generates suggested completions based on language models and context, enabling the system to assist the user without requiring explicit commands or manual intervention

Inventive Principle:
Principle #25Self-service

2Ease of operation

If handwriting recognition is used for text input, then users can write naturally as on paper, but the input process is relatively slow compared to keyboard-based methods

Engineering Contradiction:
Improvenatural writingVSAvoidinput speed
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The system merges the advantages of handwriting recognition with predictive text input. While maintaining the ability to recognize handwritten input naturally, the system combines this with language modeling and word prediction capabilities, allowing users to benefit from both natural writing input and speed-enhancing predictions

Inventive Principle:
Principle #5Merging (Combining)

3Ease of operation

If speech recognition is used for text input, then users can speak naturally, but accuracy is limited and noisy environments have adverse effects

Engineering Contradiction:
Improvenatural speechVSAvoidrecognition accuracy
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system uses text prediction as an intermediary between user input and final text output. The prediction engine acts as a mediator that refines and corrects input, whether from speech, handwriting, or typing, by suggesting contextually appropriate completions that improve accuracy while maintaining natural input methods

Inventive Principle:
Principle #24Intermediary (Mediator)

4Device complexity

If static dictionary prediction is used, then the system is simple to implement, but words sharing the same key sequence are always suggested in the same order regardless of usage

Engineering Contradiction:
Improvesystem simplicityVSAvoidprediction adaptability
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The system transitions from static to dynamic prediction by implementing language models that adapt to user behavior and context. The prediction engine dynamically adjusts suggestions based on usage frequency, contextual analysis, and user preferences, allowing the system to evolve and personalize predictions rather than following fixed sequences

Inventive Principle:
Principle #15Dynamics

5Adaptability or versatility

If dynamic dictionary with recent usage ordering is used, then the system adapts to user input patterns, but it does not keep track of which completion is the most probable

Engineering Contradiction:
Improveusage-based adaptationVSAvoidprediction accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system implements feedback mechanisms where user selections and corrections are continuously analyzed to improve prediction accuracy. The language models learn from user behavior patterns and contextual information, adjusting prediction probabilities based on feedback about which completions users actually select, thereby improving both adaptability and precision

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP2889729B1System and method for inputting text into electronic devices
Publication Date: 2023.03.15 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP2889729B1 patent drawingFigure 1
  • EP2889729B1 patent drawingFigure 2a
  • EP2889729B1 patent drawingFigure 2b

AI summary

The present invention provides a touch-screen interface that comprises: a typing pane to display inputted text; and a word prediction pane; wherein the word prediction pane is configured to present a predicted phrase.