Media Recommendation System Correcting Popularity Bias

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing media program recommendation systems are inaccurate due to popularity skewing, where popular media programs are incorrectly assumed to be related to others, leading to inappropriate recommendations.

Innovation Solution

A method and apparatus that compute a measure of implied similarity between media programs by accepting user feedback, normalizing user preferences, and correcting for popularity biases to provide more accurate recommendations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If user feedback is aggregated to determine media program similarity, then recommendation coverage is improved, but popularity bias causes inaccurate recommendations

Engineering Contradiction:

Improverecommendation coverageVSAvoidsimilarity measurement accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent transforms the raw feedback data into a normalized similarity measure by changing the parameter space from raw feedback counts to standardized similarity scores. This involves computing similarity between media programs based on normalized user feedback patterns, thereby eliminating popularity bias while maintaining comprehensive recommendation coverage.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent introduces an intermediary similarity measurement mechanism that mediates between raw user feedback and final recommendations. This intermediary layer processes and normalizes feedback data to remove popularity bias, acting as a buffer that preserves recommendation coverage while improving accuracy.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If popular media programs are used as feedback anchors, then recommendation volume is increased, but recommendation quality deteriorates

Engineering Contradiction:

Improverecommendation volumeVSAvoidrecommendation quality

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent changes the parameter used for measuring program relatedness from raw feedback volume to normalized similarity metrics. This transformation allows the system to maintain high recommendation volume while improving quality by identifying programs with similar feedback patterns rather than simply aggregating popular programs.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mechanical aggregation of popular programs with a more sophisticated feedback pattern matching mechanism. Instead of simply counting feedback volume, the system analyzes feedback patterns and computes similarity metrics, substituting a more intelligent approach that maintains volume while enhancing quality.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS8533066B2Method and apparatus for recommending media programs based on correlated user feedback

Publication Date: 2013.09.10 HULU LLC

AI summary

A method and apparatus for recommending a media program of a set of media programs to a user of a set of users is disclosed. The method and apparatus computes a measure wij of the implied similarity of a first media program (i) and a second media program (j) that corrects for the popularity of the media programs, thus resulting in a more accurate indication of the relatedness of the media programs.