Predictive Model Repository with Dynamic Retraining
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing predictive modeling systems face challenges in efficiently managing large volumes of training data and selecting effective predictive models, requiring significant computational resources and expertise, and lack the ability to dynamically update models with new data.
Innovation Solution
A computer-implemented system that stores training data and functions, generates and updates predictive models based on effectiveness scores, and provides access to the most effective model, allowing for incremental data updates and model regeneration without requiring client-side resources or expertise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional predictive modeling systems are used to manage large volumes of training data, then model training can be performed, but significant computational resources and expertise are required
Solution Approach 1:
The patent introduces a data store that acts as an intermediary between training data and predictive modeling systems. This data store pre-processes and organizes training data, reducing the computational burden on modeling systems. The data store serves as a mediator that handles data management tasks, allowing predictive models to access ready-to-use training data without requiring complex data processing infrastructure.
Solution Approach 2:
The system performs preliminary actions by pre-processing and storing training data in the data store before it is needed for model training. Training data is collected, validated, and organized in advance, so when predictive modeling needs to occur, the data is already prepared and accessible. This eliminates the need for complex real-time data processing during model training.
2Adaptability or versatility
If traditional predictive modeling systems are used, then models can be trained, but they lack the ability to dynamically update with new data
Solution Approach 1:
The patent implements dynamics by enabling predictive models to be continuously updated with new training data from the data store. The system allows models to transition from static to dynamic states, where they can incorporate new information over time. The data store continuously receives new training data that can be used to retrain and update predictive models, making the system adaptable to changing conditions without requiring complete system redesign.
Solution Approach 2:
The system ensures continuity of useful action by maintaining an ongoing process of data collection, storage, and model updating. The data store continuously accumulates training data, and predictive models can continuously access this data for updates. This creates a continuous cycle of learning and adaptation, where the system never stops improving its predictive capabilities through incremental data accumulation and model retraining.
3Adaptability or versatility
If client systems perform predictive modeling locally, then models can be trained and updated, but extensive computational resources are required at the client side
Solution Approach 1:
The patent extracts the computationally intensive data storage and management functions from client systems and places them in a centralized data store. Client systems no longer need to maintain large training data sets locally or perform complex data processing. The data store handles all data management tasks, and client systems only need to access pre-processed data for their predictive modeling needs, significantly reducing their computational resource requirements.
Solution Approach 2:
The data store serves multiple functions universally: it stores training data, validates data quality, organizes data for different predictive models, and provides data access to multiple client systems. This single universal data store replaces the need for each client system to maintain separate data infrastructure, reducing overall computational resource requirements while maintaining model customization capabilities for each client.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods, systems, and apparatus, including computer programs encoded on one or more computer storage devices, for training and retraining predictive models. Training data sets belonging to a client entity are received, e.g., over a network from a client computing system. The training data sets are used with training functions to generate trained predictive models. An effectiveness score is generated for each of the trained predictive models. A first trained predictive model is selected for the client entity from among the trained predictive models based on the respective effectiveness scores. Access to the first trained predictive model is provided to a user authorized by the client entity to access the first trained predictive model.