Enterprise Social Feed Relevance Filtering via Multi-Graph Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In cloud computing environments, conventional database management techniques fail to effectively filter and present relevant information to users in online social networks, leading to inefficiencies and timeliness issues in identifying important activities and influencers within the network.
Innovation Solution
The implementation of a system that filters content in online social networks by aggregating data from multiple graphs, including social, interest, and sentiment graphs, to identify and present relevant information sources, using highlight sources such as inner circle, global influence, topical influence, and serendipity highlights, tailored to the user's role and context.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional database management techniques are used to filter and present information in online social networks, then the system structure remains simple and easy to implement, but the relevance and timeliness of information presented to users deteriorates
Solution Approach 1:
The patent segments the information filtering system into multiple specialized graphs (social graph, interest graph, sentiment graph) that each handle specific aspects of relevance determination. This segmentation allows the system to process different dimensions of information relevance independently, improving measurement precision without overwhelming system complexity
Solution Approach 2:
The patent introduces intermediary components including a graph aggregation service and highlight source identification module that mediate between the raw database content and the user-facing information stream. These intermediaries process and filter information using multiple graphs, thereby improving relevance while isolating the complexity from the core database management system
2Loss of information
If multiple graphs (social, interest, sentiment) are aggregated to identify relevant information sources, then the relevance and timeliness of information improves, but the computational resources and processing time increases
Solution Approach 1:
The patent implements preliminary action by pre-computing and maintaining multiple specialized graphs (social, interest, sentiment) that capture different aspects of user relationships and information relevance. By having these graphs pre-established, the system avoids the need for intensive real-time computation when filtering information, thus reducing actual processing energy consumption while maintaining information completeness
Solution Approach 2:
The patent applies partial action by selectively querying and aggregating only the necessary portions of the multiple graphs based on the specific filtering context and user profile, rather than processing entire graphs. This selective approach maintains information completeness where needed while reducing overall computational resource usage
3Productivity
If the system filters and presents only relevant information from multiple graphs, then user productivity and efficiency improves, but the quantity of information available to users decreases
Solution Approach 1:
The patent applies local quality by tailoring the information filtering and presentation to each user's specific context, role, and preferences. Different users receive different subsets of information from the multiple graphs based on their local needs and characteristics, thereby maximizing productivity for each user while preserving the full quantity of information in the system for those who need it
Data Source
AI summary
Disclosed are techniques for identifying data items to display in a feed of an enterprise social networking system. A first subset of data items is identified as associated with an inner circle of data sources with which a user interacts at or exceeding a designated frequency, a second subset of data items is identified as associated with an interest graph associated with the user, and a third subset of data items is identified as associated with a group or an organization. A highlights feed can be displayed with different channels, where a first channel is defined by the first subset of data items, a second channel is defined by the second subset of data items, and a third channel is defined by the third subset of data items.