Bipartite Graph Recommendation Indicator Determination

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing recommendation systems suffer from low performance due to the 'ignoring' or 'blacklist' processing mechanisms for negative feedback information, which fail to comprehensively utilize both positive and negative feedback features.

Innovation Solution

A method involving the construction of two bipartite graphs based on interaction feature data, one for positive feedback and one for negative feedback, to determine comprehensive embedding vector representations of graph nodes, which are then used to calculate recommendation indicators for resource information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If negative feedback information is processed using 'ignoring' or 'blacklist' mechanism, then the system complexity is reduced and ease of operation is improved, but the recommendation performance and accuracy deteriorate

Engineering Contradiction:

Improveease of operationVSAvoidrecommendation accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The patent segments the feedback information processing into two distinct bipartite graphs: one for positive feedback and one for negative feedback. This segmentation allows the system to handle different types of feedback separately while maintaining comprehensive analysis, resolving the contradiction by structuring complexity in a manageable way that improves accuracy without sacrificing operational simplicity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent merges positive and negative feedback information into a unified recommendation framework by constructing two bipartite graphs that are then integrated through graph neural network processing. This merging allows comprehensive utilization of all feedback types, improving recommendation accuracy while the automated processing maintains ease of operation.

Inventive Principle:

Principle #5Merging (Combining)

2Measurement precision

If both positive and negative feedback information are comprehensively utilized through dual bipartite graphs, then recommendation accuracy is improved, but the device complexity increases

Engineering Contradiction:

Improverecommendation accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides the complex feedback processing task into two separate bipartite graphs (positive feedback graph and negative feedback graph), making the overall system more manageable. Each graph handles a specific type of feedback, reducing the complexity of individual components while maintaining comprehensive analysis capabilities.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces graph neural networks as an intermediary mechanism that automatically processes and integrates information from both bipartite graphs. This intermediary handles the complexity of combining positive and negative feedback, reducing the need for manual system configuration and lowering operational complexity despite the comprehensive analysis performed.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Loss of information

If negative feedback information is discarded or blacklisted, then the loss of information is reduced, but the recommendation performance deteriorates

Engineering Contradiction:

Improveinformation retentionVSAvoidrecommendation performance

Core Design Contradiction:

Loss of informationVSProductivity

Solution Approach 1:

The patent converts negative feedback information, which was previously discarded or blacklisted, into a beneficial resource by incorporating it into a dedicated bipartite graph. The graph neural network processes this negative feedback to improve recommendation accuracy, transforming what was once harmful or useless information into a valuable input that enhances system performance.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent implements a feedback mechanism where both positive and negative user interactions are captured, processed through graph neural networks, and used to continuously improve recommendation indicators. This closed-loop feedback system ensures that all user feedback, including negative feedback, contributes to enhancing recommendation performance over time.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20250181890A1Determining recommendation indicator of resource information

Publication Date: 2025.06.05 TENCENT TECHNOLOGY (SHENZHEN) CO LTD

AI summary

In a method for determining recommendation indicators, a first bipartite graph is constructed based on interaction feature data that indicate interactions between a plurality of objects and a plurality of pieces of resource information. A second bipartite graph is constructed based on the interaction feature data. Comprehensive embedding vector representations of the plurality of graph nodes are determined based on the first bipartite graph and the second bipartite graph. The recommendation indicator for each of the plurality of pieces of resource information with respect to each of the plurality of objects is determined based on the comprehensive embedding vector representation of the object node corresponds to the respective object and the comprehensive embedding vector representation of the resource information node corresponds to the respective piece of resource information. Apparatus and non-transitory computer-readable storage medium counterpart embodiments are also contemplated.