ML Conversation Graph for Personalized Content Delivery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in efficiently generating personalized conversation flows and selecting relevant digital content for users in online services due to the large volume of user data, leading to excessive computational resource consumption.
Innovation Solution
A machine learning system generates a conversation graph using user data to create a personalized conversational flow, dynamically adjusting based on user interactions and selecting relevant digital content through a job-seeker coach bot, which includes nodes and edges representing interactive dialogues and transitions, and utilizes embeddings to simplify similarity analysis among users and job content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If personalized search analyzes user characteristics against a corpus of possible results to find the best options, then search relevance and user engagement are improved, but computational resource consumption increases excessively
Solution Approach 1:
The patent segments the user population into clusters based on similar characteristics and behaviors. Instead of analyzing every user against the entire corpus individually, the system processes clustered groups of users together, reducing the computational complexity while maintaining personalized search relevance through cluster-level analysis
Solution Approach 2:
The system performs preliminary clustering of users based on their characteristics before conducting the actual search analysis. This pre-grouping action reduces the search space and computational resources needed during the main search operation, as the system only needs to analyze within clusters rather than across the entire user population
2Adaptability or versatility
If the system processes large amounts of user data to find similarities among users, then personalized content delivery is improved, but electronic resource consumption increases
Solution Approach 1:
The patent divides the large user dataset into manageable clusters of similar users. This segmentation allows the system to process and analyze user data in smaller, organized groups rather than handling the entire large dataset at once, reducing electronic resource consumption while maintaining the ability to provide personalized content through cluster-level patterns
Solution Approach 2:
The system changes the approach from individual user analysis to cluster-level analysis by transforming the data representation. This parameter change allows the system to capture personalized information through aggregate cluster characteristics, reducing the computational resources needed while maintaining personalization effectiveness
Data Source
AI summary
A machine for improving content delivery generates a graph representing a personalized conversational flow for sequenced delivery of digital content. The graph includes nodes representing interactive dialogues between a machine and a user, and edges that connect the nodes. The machine causes display of a user interface including a prompt related to job-seeking guidance. The machine, based on a first action in response to the prompt, dynamically adjusts the graph, the dynamic adjusting including selecting a first node. The machine generates and causes display of a first incentive content item, and a first call-to-action content item. The machine, in response to a second action received in response to the first call-to action content item, dynamically selects an edge connecting the first node and a further node. The dynamic selecting of the edge results in display of a further incentive content item, and a further call-to-action content item.