Soft Input Panel String Prediction with Confidence Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current soft input panels on mobile devices are limited in predicting phrases or sentences, fail to consider user typing history, and often overwhelm users with too many distracting suggestions, lacking an undo feature to correct accidental selections.
Innovation Solution
A system that analyzes user history to provide suggested phrases before character input, limits displayed predictions by confidence value, and includes an undo feature to correct inadvertent selections, using a candidate prediction generator and selector to refine and display string predictions based on historical data and user preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the soft input panel displays multiple predictions to cover more possibilities, then the completeness of predictions is improved, but the user distraction increases and the correct prediction becomes harder to select
Solution Approach 1:
The patent applies partial action by displaying only the top N predictions with highest confidence values rather than all possible predictions. This selective display maintains sufficient prediction coverage while avoiding user distraction from low-confidence options, directly resolving the contradiction between prediction completeness and selection ease.
Solution Approach 2:
The patent changes the parameter of prediction ranking by using confidence values to order predictions. By sorting predictions based on their confidence scores and displaying only those above a threshold, the system achieves both comprehensive coverage of likely predictions and ease of selection through clear prioritization.
2Adaptability or versatility
If the soft input panel displays less likely predictions to provide more options, then the versatility of predictions is improved, but the user distraction increases
Solution Approach 1:
The patent extracts only the most relevant predictions from the full set of possible predictions based on confidence values. By taking out and displaying only the top N highest-confidence predictions, the system eliminates distracting low-confidence options while maintaining versatile prediction coverage for the most likely user intentions.
Solution Approach 2:
The patent applies partial action by providing only the necessary prediction options (top N by confidence) rather than all possible predictions. This partial display approach maintains sufficient versatility for common use cases while removing the harmful distraction of unlikely predictions.
3Speed
If the soft input panel provides predictions based only on current input characters, then the responsiveness is improved, but the accuracy of predictions decreases
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing prediction data based on historical input patterns and context before the user actually types. The system maintains a database of predicted responses mapped to trigger phrases, allowing instant retrieval and display of accurate predictions without waiting for extensive user input, thus achieving both responsiveness and accuracy.
Solution Approach 2:
The patent uses feedback from historical user behavior data to continuously improve prediction accuracy. By analyzing past typing patterns and preferred responses, the system refines its prediction models to provide more accurate suggestions that respond appropriately to current input context, resolving the contradiction between speed and precision.
4Device complexity
If the soft input panel does not provide an undo feature, then the simplicity is improved, but the user frustration increases upon mis-selection
Solution Approach 1:
The patent applies beforehand cushioning by implementing an undo feature that prepares for and mitigates the harmful effect of mis-selections. By allowing users to undo incorrect predictions, the system cushions against user frustration and error without significantly increasing interface complexity, as the undo functionality builds upon existing selection mechanisms.
Data Source
AI summary
In a mobile device, the text entered by users is analyzed to determine a set of responses commonly entered by users into text applications such as SMS applications in response to received messages. This set of responses is used to provide suggested responses to a user for a currently received message in a soft input panel based on the text of the currently received message. The suggested responses are provided before any characters are provided by the user. After the user provides one or more characters, the suggested responses in the soft input panel are updated. The number of suggested responses displayed to the user in the soft input panel is limited to a total confidence value to reduce user distraction and to allow for easier selection. An undo feature for inadvertent selections of suggested responses is also provided.