Dynamic Relevance Scoring for New Information Cards
Solution Overview
Problem
Existing computer user interfaces struggle to effectively display new available information due to limited screen space and the challenge of determining relevance, often failing to present relevant information as models are trained based on previous user interactions, leading to new information being overlooked.
Innovation Solution
An electronic device process that detects new available information, increases its relevance score above a display threshold, displays it, and adjusts the score based on user interaction data, ensuring relevant information is presented and allowing for model training on user behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a model trained on previous user interactions is used to determine relevance of information, then the model can accurately determine relevance of established information types, but it performs poorly on new available information with limited or no interaction data
Solution Approach 1:
The system proactively displays new available information by increasing its relevance score above the display threshold before sufficient user interaction data is collected. This preliminary action allows the system to present new information types to users even though the model lacks historical interaction data for accurate relevance determination, thereby enabling model training on new information types while maintaining accurate display of both new and established information
Solution Approach 2:
The system collects user interaction data with displayed new information and uses this feedback to train and update the relevance determination model. By continuously incorporating user feedback, the model improves its ability to accurately determine relevance for new information types over time, resolving the contradiction between current accuracy and future adaptability
2Loss of information
If all available information is displayed in the user interface, then complete information availability is achieved, but screen space is exceeded and users become overwhelmed
Solution Approach 1:
The system applies different display treatments to different types of information based on their relevance characteristics. New available information receives enhanced treatment by being displayed with increased relevance scoring, while established information types are displayed according to their normal relevance-based filtering. This local differentiation allows the system to prioritize new information without displaying all available information, maintaining usability while improving information completeness
Solution Approach 2:
The system dynamically adjusts the relevance score parameter for new available information, increasing it above the display threshold to ensure these items are included in the displayed subset. This parameter modification allows the system to override the normal relevance-based filtering for new information, ensuring both new and established information are appropriately represented in the limited display space
Data Source
AI summary
An example method includes, at an electronic device with one or more processors and memory: detecting a new type of available information, the new type of available information being associated with a card for display in a content feed; in accordance with detecting the new type of available information, increasing a relevance score of the card above a display threshold; in accordance with increasing the relevance score above the display threshold: causing the card to be displayed in the content feed; and obtaining data representing user interaction with the displayed card; adjusting, based on the data representing user interaction with the displayed card, the relevance score; and modifying a display of the card in the content feed based on the adjusted relevance score.