Context Feature Model for Privacy-Preserving Ad Placement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of poor advertisement placement due to the increasing difficulty in obtaining user identities and profiles, resulting from enhanced user information protection, leads to a decline in advertisement effectiveness.
Innovation Solution
A model training method that utilizes context features to replace user features, incorporating an auxiliary network to enhance interaction modeling between context and advertisement features, thereby improving representation and enabling precise advertisement placement without relying on user profiles or behaviors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If user information protection is enhanced to protect privacy, then user privacy protection is improved, but obtaining user identity and profile information becomes more difficult, resulting in poor advertisement placement effect
Solution Approach 1:
The patent introduces context features as an intermediary to replace direct user profile information. Instead of relying on user identity and behavior data, the system uses context features (page content, device information, location, time) as mediators to infer user preferences and perform advertisement placement, thereby protecting user privacy while maintaining advertising effectiveness
Solution Approach 2:
The patent replaces the traditional mechanical approach of directly collecting and using user profile data with a new mechanism based on context feature extraction and interaction modeling. By substituting user feature-based models with context feature-based models, the system achieves advertisement placement without directly accessing user information
2Reliability
If traditional user feature-based models are used for advertisement placement, then advertisement placement effect can be maintained, but dependence on user feature information increases, conflicting with privacy protection requirements
Solution Approach 1:
The patent inverts the traditional advertisement placement approach by switching from user-feature-centered modeling to context-feature-centered modeling. Instead of starting with user profiles and matching them to advertisements, the system starts with context features and uses interaction modeling to infer advertisement relevance, fundamentally reversing the information flow and dependency structure
3Reliability
If context features are used to replace user features, then dependence on user feature information is reduced, but new interaction modeling between context and advertisement features is required, increasing model complexity
Solution Approach 1:
The patent segments the feature processing into distinct components: context feature extraction, advertisement feature extraction, and interaction modeling. By dividing the model into separate modules (context tower, advertisement tower, and interaction layer), the system manages complexity through modular architecture while achieving effective privacy-protected advertisement placement
Data Source
AI summary
An advertisement placement method is provided, and may be applied to an electronic device having a display. The method includes: obtaining a first access operation of a user on a first page; in response to the first access operation, obtaining a first context feature related to the first page, where the first context feature is irrelevant to a user profile and/or a user behavior of the user; and displaying f advertisements based on the first context feature, where f≥1. In this way, a context feature is used to replace a user feature, to reduce dependence on the user feature, and implement precise advertisement placement when a feature, for example, a user profile, cannot be obtained.