Identity Graph Enhancement via Probabilistic ML Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advertisers face limitations in reaching a broader audience and evaluating content-related communications using identity graphs solely based on deterministic information, which restricts the effectiveness of advertising efforts in real-time bidding frameworks.
Innovation Solution
Enhancing deterministic identity graphs with probabilistic data by using communication-related information to identify probable locations for user devices, generating nodes and edges based on association thresholds, and mapping identifiers to facilitate content delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If identity graphs are built solely on deterministic information, then data reliability is maintained, but the ability to reach broader audiences and evaluate content-related communications is limited
Solution Approach 1:
The patent merges deterministic identity graph data with probabilistic data from machine learning models to create an enhanced identity graph. This combination allows the system to maintain the reliability of deterministic data while adding the versatility to reach broader audiences by incorporating inferred user attributes and relationships that were previously inaccessible.
Solution Approach 2:
The patent introduces machine learning models as intermediaries between deterministic identity data and the broader user population. These models process deterministic data and generate probabilistic inferences that bridge the gap between known user identities and potential target audiences, enabling expanded reach without completely sacrificing data reliability.
2Adaptability or versatility
If probabilistic data is added to identity graphs, then the ability to target content is improved, but system complexity increases
Solution Approach 1:
The patent segments the identity graph into deterministic nodes and probabilistic nodes, with clear differentiation between the two types of data. This segmentation allows the system to manage complexity by processing and storing different types of data in separate, organized structures, while still enabling sophisticated content targeting through their interaction.
Solution Approach 2:
The patent changes the parameters of the identity graph system by introducing confidence scores, probability weights, and threshold values that govern how probabilistic data is integrated. These parameter changes provide control mechanisms that manage system complexity by allowing configurable integration of probabilistic data without overwhelming the deterministic framework.
3Measurement precision
If machine learning models are used to infer user locations, then content delivery accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary actions by pre-processing deterministic user data and training machine learning models in advance. The models are trained on historical data to recognize patterns and make inferences about user locations and preferences. This preliminary training allows the system to make accurate location inferences quickly during actual content delivery without requiring real-time complex computations.
Solution Approach 2:
The patent applies partial action by using machine learning models to infer only the specific attributes needed for content targeting, rather than processing all possible user data. The models are designed to make targeted inferences about location and relevant user characteristics, reducing the computational burden while maintaining sufficient accuracy for effective content delivery.
Data Source
AI summary
Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for enhancing a deterministic identity graph with probabilistic data. An example embodiment operates by identifying a node for a location indicated by an identity graph. Receiving user device information based on an indication that a user device is within proximity to the location. Generating a node for the user device on the identity graph based on the indication of the user device satisfying an association threshold. Generating an edge between the node for the location and the node for the user device based on a weighted value for an attribute of the user information. Mapping an identifier for the user device to an identifier of the location based on a distance of the edge and causing a content item to be sent to the user device based on the identifier mapping.