Notification Filtering Using Machine Learning Relevancy Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users are overwhelmed with numerous notifications from various applications on their devices, making it impractical to act on each one, leading to 'notification blindness' and missed important information.
Innovation Solution
A system that uses machine learning models to determine a user's active context and filter notifications based on historical behavior, presenting only relevant notifications at optimal times, while allowing service notifications to be prioritized and time-sensitive notifications to be displayed before expiration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the system presents all notifications from various applications, then users receive complete information, but users experience notification blindness and cannot find important information
Solution Approach 1:
The notification system segments notifications into different categories (important vs. non-important) based on machine learning analysis of user behavior patterns. The ML model divides the notification stream into priority levels, presenting only relevant notifications to users while filtering out less important ones, thus preventing notification blindness while preserving access to critical information.
2Ease of operation
If the system filters out notifications based on relevance, then users experience less annoyance, but users may miss important notifications
Solution Approach 1:
The system implements feedback loops where user interactions with notifications (opening, dismissing, ignoring) are continuously monitored and fed back to the machine learning model. This feedback refines the relevance scoring algorithm, improving its ability to accurately distinguish important notifications from non-important ones over time, thereby maintaining high notification delivery accuracy while reducing user annoyance.
Solution Approach 2:
The notification filtering system is dynamic and adaptive, continuously adjusting its filtering criteria based on changing user behavior patterns and context. The machine learning model evolves its understanding of what constitutes an important notification for each user, allowing the system to maintain high reliability in notification delivery while adapting to user preferences and reducing annoyance as conditions change.
3Speed
If the system presents notifications immediately when generated, then users receive timely information, but users are overwhelmed by the volume of notifications
Solution Approach 1:
The system performs preliminary analysis of notifications using machine learning models before presenting them to users. The ML model pre-evaluates each notification's relevance based on user behavior patterns, app importance, and contextual factors, filtering out non-important notifications before they reach the user. This preliminary action reduces the quantity of notifications users see while maintaining timely delivery of important ones.
4Measurement precision
If the system uses machine learning models to analyze user behavior, then notification relevance improves, but system complexity increases
Solution Approach 1:
The machine learning model serves multiple functions within the notification system: it scores notification relevance, predicts user engagement likelihood, adapts to changing user preferences, and prioritizes notifications for display. This multi-functionality reduces the need for separate complex systems while maintaining high measurement precision in notification relevancy assessment.
Data Source
AI summary
Disclosed embodiments pertain to filtering application notifications from one or more applications on an electronic device. A request can be received from an application installed on the electronic device to present a notification. In response, a machine learning model is invoked that is trained to generate a relevancy score for a notification based on historical behavior of the user with respect to previous notifications from the application as well as the active context of the user. The relevancy score can be compared with a threshold to determine whether the threshold is satisfied. Notification filtering can be performed based on whether or not the threshold is satisfied. The method can filter out the notification from presentation if the threshold is satisfied and present the notification if the threshold is unsatisfied, or vice versa.