Inference Model for Shared Device User Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In scenarios where a user device is shared among multiple users, existing systems face challenges in determining the identity of the specific user interacting with applications, leading to inefficient content delivery and resource utilization, as content may be sent to users who are not interested, wasting network, storage, and processor resources.
Innovation Solution
An online system employs an inference model trained with machine learning on historical usage data to determine the most likely user associated with a device based on device usage parameters, historical data, and activity indicators, using an identity graph to link user accounts and devices, and selecting content items tailored to the inferred user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If content is delivered based on device identifier without user identification, then content delivery can proceed without additional processing, but content relevance to the actual user deteriorates and resources are wasted
Solution Approach 1:
The system implements feedback loops where user interactions with content are continuously monitored and fed back into the inference model. This feedback refines the model's ability to predict current user identity, making subsequent content delivery more accurate without requiring additional identification steps, thus improving both relevance and resource efficiency
Solution Approach 2:
The inference model performs preliminary user identification based on historical data and device usage parameters before content delivery occurs. This preliminary action establishes user identity context in advance, allowing the system to deliver relevant content without requiring additional real-time processing or user input, thereby maintaining delivery speed while improving relevance
2Measurement precision
If an inference model is implemented to identify users on shared devices, then content relevance to the actual user is improved, but system complexity increases
Solution Approach 1:
The inference model acts as an intermediary layer between device identification and user profile matching. Rather than directly linking devices to users, the model processes device usage parameters and historical data to infer current user identity. This intermediary approach maintains system modularity while improving identification accuracy without requiring fundamental architectural changes
Solution Approach 2:
The system creates simplified copies of user identification logic through the inference model, which replicates the decision-making process of user identification based on observable device usage patterns. This copying approach allows the system to handle user identification without requiring direct access to complex user authentication mechanisms, thereby improving accuracy while managing complexity
3Reliability
If content is sent to all users of a shared device, then no user is left without potentially relevant content, but resource utilization deteriorates due to sending content to uninterested users
Solution Approach 1:
The system dynamically changes the parameter of user identity from a static device identifier to a dynamic inferred user identity based on real-time device usage parameters. By adjusting this parameter, the system can selectively deliver content to the actual current user rather than all potential users of the device, maintaining content coverage while dramatically improving resource utilization efficiency
Data Source
AI summary
An online system receives a request from a third-party application for a content item to be delivered to a user of a user device, the request including a device identifier corresponding to the user device used to access the third-party application. In order to deliver a targeted content item, the online system accesses inference model to infer an identity of the user, based upon device usage parameters based upon the received request, historical data associated with the device identifier, or an indication of activity associated with the device identifier on the online system or one or more first-party applications.