Contextual Content Insertion Using Privacy-Safe Tag Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in delivering supplemental content effectively due to diminishing capabilities for matching viewer personal data and tracked activity, constrained by regulations and platform terms, leading to inefficiencies in reaching intended audiences.
Innovation Solution
A system utilizing machine learning for contextual tagging to generate tags for main content, enabling contextual matching of supplemental content without relying on user-specific data, and optimizing insertion based on context analysis and affinity graphs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If viewer personal data and tracked activity are used for supplemental content matching, then viewer engagement and targeting precision are improved, but regulatory compliance and system reliability deteriorate due to intensifying regulations and platform terms
Solution Approach 1:
The patent extracts and removes personal data and tracked activity from the supplemental content matching process. Instead of using viewer-specific data, the system uses only contextual information from the main content itself (video metadata, audio metadata, text metadata) to generate contextual tags and match supplemental content, thereby complying with regulations while maintaining matching capability
Solution Approach 2:
The patent introduces contextual tags as an intermediary between the main content and supplemental content. These tags are generated from contextual analysis of the main content without requiring personal data, serving as a mediator that enables relevant matching while adhering to privacy regulations and platform terms
2Productivity
If contextual tagging and analysis are implemented for supplemental content insertion, then content delivery speed and efficiency are improved, but storage requirements and system complexity increase
Solution Approach 1:
The patent performs contextual tagging and analysis in advance, before supplemental content insertion is needed. Contextual tags are generated from main content metadata during content ingestion or preprocessing, so that when supplemental content needs to be inserted, the matching can occur quickly using pre-computed tags without real-time analysis complexity
Solution Approach 2:
The patent segments the contextual analysis into distinct metadata components (video metadata, audio metadata, text metadata) and processes each separately to generate contextual tags. This segmentation simplifies the overall system by breaking down the complex analysis task into manageable, independent modules that can be processed separately
Data Source
AI summary
In some embodiments, a method stores a set of tags from a taxonomy for a break in main content. A portion of the main content within a time period threshold of the break is analyzed to determine the set of tags. An indication of the break that is going to be experienced during playback of main content is received. A client device is playing back the main content. Responsive to the indication, the set of tags for the break is retrieved. The method provides information for the set of tags to a supplemental content system to facilitate selection of an instance of supplemental content based on the set of tags. The method provides information for the instance of supplemental content to the client device to insert the instance of supplemental content in the break during the playback of the main content.