Graph Database Media Recommendation and Dynamic Micro-Licensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional recommender systems in digital media distribution face challenges in providing personalized recommendations due to the 'cold start problem' and non-technical restrictions, leading to suboptimal user experiences and cost/benefit ratios, as they require extensive user data for accurate predictions and impose limitations on interactive access to media content.
Innovation Solution
A client/server-based digital media distribution system that utilizes a graph database to analyze historical usage data and generate recommendations by identifying co-occurrences between media works, allowing for personalized content suggestions with minimal user data requirements and flexible interaction options through a dynamic micro-licensing transaction and interaction entitlement system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional recommender systems use collaborative filtering or content-based filtering techniques, then they can generate recommendation lists, but they require extensive user data which creates the 'cold start problem' and reduces effectiveness for new users
Solution Approach 1:
The patent introduces an intermediary entity (the online media provider or service) that mediates between users and media works. This intermediary observes and records user interactions with media content, using this observed behavior data to generate recommendations without requiring extensive explicit user input or profiles, thereby solving the cold start problem while maintaining recommendation accuracy
Solution Approach 2:
The system enables users to implicitly provide recommendation data through their natural media consumption behavior. Users' interactions with media works (viewing, listening, engaging) automatically generate the data needed for recommendation algorithms, eliminating the need for users to manually input preferences or complete extensive profiles
2Reliability
If online media providers impose non-technical restrictions on media access, then they can control licensing and distribution, but they limit user interaction with media content and reduce user experience quality
Solution Approach 1:
The patent implements dynamic licensing restrictions that adapt based on user behavior, preferences, and engagement patterns. Rather than static access controls, the system dynamically adjusts media interaction permissions, allowing users to earn or gain access to additional interaction types through their usage patterns, thereby balancing licensing control with user experience flexibility
Solution Approach 2:
The system changes the parameters of media access restrictions based on observed user behavior. Licensing parameters such as interaction types, access duration, and media selection freedom are dynamically adjusted according to user engagement metrics, transforming rigid licensing controls into flexible, behavior-adaptive access policies
3Device complexity
If conventional systems require users to choose between non-interactive streaming and interactive access, then they can simplify licensing models, but they force users into suboptimal cost/benefit ratios and reduce personalization
Solution Approach 1:
The patent segments the traditional binary licensing model into multiple granular access levels and interaction types. Instead of forcing users into broad categories of non-interactive or interactive access, the system divides media access into finer segments (different interaction types, time periods, media selections) that can be individually licensed and adjusted based on user behavior and preferences
Data Source
AI summary
Provided herein are systems and methods for generating a set of recommended media works based on an identified seed media work. The seed media work is used to identify a set of co-occurring media works, which are defined as media works that have been intentionally interacted with within a predetermined time window as the seed media work by a single user. After the co-occurrence set is created, it is ranked and filtered. Media works identified in the co-occurrence set may then be selected based on their ranking.