User-guided term suggestions via cursor activity analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional spell-correction systems fail to provide relevant suggestions for misspelled terms as they do not consider user-initiated activities near the misspelled terms, leading to less relevant replacement options.

Innovation Solution

A method and apparatus that identify user-initiated activities near a given term to provide suggested terms, such as spell-corrected, modified, or alternative terms, by determining an edit position and selecting candidate terms based on similarity measures, relevance scores, and historical edits, thereby increasing the relevance of suggested terms to the user's activity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional spell-correction systems provide suggestions based only on mapping between misspelled terms and corrections, then the system is simple to implement, but the relevance of suggested terms to user intent is low

Engineering Contradiction:
Improverelevance of suggested termsVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary analysis of user-initiated activity near the misspelled term before generating corrections. By detecting cursor position, selection state, and surrounding context in advance, the system prepares relevant suggestions that match user intent, resolving the contradiction between simplicity and relevance.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses user-initiated activity (cursor position, text selection) as feedback to adjust which corrections are suggested. This feedback loop allows the system to adapt suggestions to user intent dynamically, improving relevance without requiring fundamentally complex architecture.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If the system considers user-initiated activity near the term, then the relevance of suggested terms improves, but the computational requirements and processing time increase

Engineering Contradiction:
Improverelevance of suggested termsVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs partial analysis by focusing only on user-initiated activity near the misspelled term rather than analyzing the entire document. This selective approach maintains high relevance while reducing computational overhead and processing time.

Inventive Principle:
Principle #16Partial or excessive action

3Reliability

If the system provides spell-corrected suggestions based on user-initiated activity, then the accuracy of suggestions improves, but the device complexity increases

Engineering Contradiction:
Improveaccuracy of suggestionsVSAvoidsystem architecture complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system uses a multi-functional approach where the same activity detection mechanism serves multiple purposes: identifying user intent, determining context, and guiding correction selection. This universal mechanism improves accuracy without proportionally increasing system complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS10459957B2User-guided term suggestions
Publication Date: 2019.10.29 GOOGLE LLC
  • US10459957B2 patent drawing
  • US10459957B2 patent drawing
  • US10459957B2 patent drawing

AI summary

Methods and apparatus related to providing user-guided term suggestions. Some implementations may be directed to identifying user input that includes at least one term and identifying a user-initiated activity near the term. An edit term may be identified based on the user-initiated activity near the term, an edit position in the edit term may be identified, and one or more candidate terms may be identified based on the edit term. Similarity measures for the candidate terms may be determined. The similarity measure of a given candidate term of the candidate terms may be based on the edit position. One or more of the candidate terms may be selected as suggested terms based on the similarity measures and the selected suggested terms provided for potential replacement of the edit term.