Shared Memory Parameter Vector Adjustment for Risk Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine learning models face challenges in accurately predicting entity risk levels due to inadequate training data structures that fail to effectively segment and map risk-scoring components from physical and behavioral health attributes, leading to suboptimal performance in identifying and ranking entities based on risk.
Innovation Solution
A data processing apparatus and method that segment attributes into distinct risk-scoring components, map them to respective categories, and generate structured training data structures to improve the machine learning model's capability to predict entity risk levels by adjusting parameter vectors based on error analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional training data structures are used without segmentation, then the data processing is simpler, but the machine learning model's prediction accuracy for entity risk levels deteriorates
Solution Approach 1:
The patent segments training attributes into distinct risk-scoring components (physical health attributes, behavioral health attributes, demographic attributes) and maps them to separate categories. This segmentation enables the machine learning model to process structured data more effectively, improving prediction accuracy for entity risk levels while maintaining manageable complexity through systematic organization.
2Productivity
If parameter adjustments are performed using traditional memory access methods, then the system architecture is simpler, but the training time and latency increase
Solution Approach 1:
The patent merges the parameter vector storage with the training data structure in shared memory, allowing the parameter vector adjustment engine to access and modify parameters during training without external memory access. This integration eliminates additional memory access latency and accelerates the training process by enabling direct in-memory parameter updates.
3Reliability
If comprehensive risk-scoring components are collected from multiple sources, then the prediction capability is improved, but the data processing complexity and time increase
Solution Approach 1:
The patent segments comprehensive training attributes from multiple sources into distinct risk-scoring components (physical health, behavioral health, demographic) and maps them to structured categories. This segmentation organizes complex multi-source data into manageable segments, improving prediction reliability while maintaining processing efficiency through systematic structure.
Solution Approach 2:
The patent transforms unstructured or semi-structured training attributes into structured parameter vectors with defined categories and risk-scoring components. This parameter transformation enables the machine learning model to efficiently process comprehensive data from multiple sources while maintaining manageable complexity through standardized data structures.
4Measurement precision
If structured training data with segmented risk-scoring components is implemented, then the machine learning model's prediction accuracy improves, but the data preparation process becomes more complex
Solution Approach 1:
The patent segments training attributes into distinct risk-scoring components and maps them to structured categories, creating a systematic framework for data preparation. While this segmentation improves prediction accuracy, the patent addresses the increased complexity by providing automated mapping mechanisms and structured data formats that streamline the preparation process.
Data Source
AI summary
Methods, systems, and apparatus for training model parameters stored in shared memory to predict risk. The method may include obtaining training data that includes a plurality of training data structures that each represent attributes of an entity, wherein each training data structure represents (i) features derived from a first set of categories defined by a first model and from a second set of categories defined by a second model, and (ii) a risk-level associated with the entity. For each respective training data structure, providing the training data structure as an input to the model, receiving an output from the model based on the model's processing of the training data structure, determining an amount of error between the output of the model and the risk-level of the training data structure, and adjusting a parameter value of the model stored in a shared memory based on the determined error.