Partial Content Metadata Linking for Accurate User Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current recommendation systems inaccurately infer user preferences due to incomplete data, as they assume user enjoyment of an entire content item based on feedback given during a partially watched section, potentially missing critical aspects such as opening mature content or disliked actors, leading to incorrect content recommendations.
Innovation Solution
A method that links metadata to user behavior information based on engagement with specific parts of content, weighting recent feedback more heavily to create a more accurate user profile, allowing for improved content and advert selection that aligns with the user's actual preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user feedback is used to update user profile, then recommendation accuracy improves, but incorrect generalizations occur when user watched only part of content
Solution Approach 1:
The patent segments user feedback into content-specific portions by tracking which specific scenes or segments of content a user has watched. Instead of treating feedback on entire content items, the system divides feedback into granular segments corresponding to actual viewing portions, preventing incorrect generalizations from partial viewing.
Solution Approach 2:
The patent applies local quality by associating user feedback with specific metadata segments rather than applying uniform feedback across entire content items. Each metadata element is weighted based on the user's actual engagement with its corresponding content portion, allowing different parts of the user profile to have different feedback weights.
2Measurement precision
If scene-level metadata is used to improve recommendation detail, then navigation precision improves, but user preference inference becomes more error-prone
Solution Approach 1:
The patent performs preliminary action by pre-associating metadata with specific content segments before user feedback is processed. This allows the system to later accurately map user feedback to specific metadata elements, ensuring that preference inference is based only on actually viewed portions rather than making errors from unseen content.
Solution Approach 2:
The patent implements feedback by continuously refining the association between user behavior and metadata based on actual viewing patterns. The system uses feedback loops to adjust which metadata elements are weighted based on user engagement, improving the reliability of preference inference over time while maintaining scene-level detail.
3Measurement precision
If user feedback is weighted by time proximity, then recent preferences are captured better, but data processing complexity increases
Solution Approach 1:
The patent applies parameter changes by introducing time proximity as a weighting parameter for user feedback. Recent feedback is weighted more heavily than older feedback, allowing the system to capture evolving user preferences dynamically. This parameter adjustment is applied systematically to metadata associations based on timing information.
Data Source
AI summary
There is disclosed a method comprising: receiving one or more items of metadata associated with an item of content, each of the items of metadata identifying a characteristic of a part of the item of content; receiving one or more items of user behaviour information corresponding to an engagement of a user with one or more parts of the item of content; and linking the metadata to the user behaviour in dependence on user engagement for the part of the content associated with the metadata.