Context-Aware Confidence Scoring for Accidental Send Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional spell check features in information handling devices often fail to prevent users from sending messages with errors due to accidental actions, such as unintentionally hitting the 'send' button, especially when autocorrect suggestions are located close to the 'send' button or when users are in a hurry and do not have time to review their messages.
Innovation Solution
Assigning a confidence score to elements like the 'send' button based on input context, such as application type, input type, input location, and grammar, and adjusting this score based on the presence of errors in the message, thereby requiring a higher confidence threshold to perform the action, preventing accidental sending of error-filled messages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional spell check features are used, then users can identify errors in messages, but users may still accidentally send messages with errors when in a hurry or when autocorrect suggestions are close to the send button
Solution Approach 1:
The patent introduces an intermediary confidence score mechanism between the user's send action and the actual message transmission. This intermediary evaluates multiple contexts (grammar, autocorrect state, input method, application type) to determine whether to allow the send action, thereby preventing accidental error-filled messages without requiring manual review from users in a hurry
Solution Approach 2:
The system provides feedback by adjusting the confidence score based on detected errors and contextual factors. When the confidence score falls below the threshold due to detected errors, the system feedback prevents the send action, informing users indirectly through the blocked transmission that their message contains issues needing correction
2Reliability
If the system prevents sending messages with errors by requiring review, then message quality improves, but users lose time reviewing and correcting messages
Solution Approach 1:
The system performs self-service by automatically evaluating the message quality through confidence score calculation based on grammar checks, autocorrect status, and contextual analysis. This automated quality assessment and prevention mechanism eliminates the need for users to manually review and correct messages, maintaining high message quality without requiring user time investment
3Ease of operation
If autocorrect suggestions are placed close to the send button for ease of access, then user operation becomes faster, but users may unintentionally hit the send button instead of autocorrect suggestions
Solution Approach 1:
The system applies preliminary anti-action by proactively analyzing the input context and message quality before the user completes the sending action. When potential errors are detected through grammar checks or improper autocorrect usage, the confidence score is reduced below the threshold, preventing the send action before the user can accidentally transmit an error-filled message
Data Source
AI summary
One embodiment provides a method, including: receiving, at an information handling device, an indication to perform an action associated with an element in an application; adjusting, using a processor, a confidence score of the element based upon at least one input context; responsive to the confidence score being higher than a predetermined threshold, performing the action; and responsive to the confidence score being lower than the predetermined threshold, performing a second action. Other aspects are described and claimed.