Hierarchical Models for Accurate Intercorrelated Synthetic Datasets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating synthetic datasets that are intercorrelated fail to accurately reproduce correlations between datasets, leading to inefficiencies and resource wastage due to the need for large, brittle models that require frequent architectural changes.
Innovation Solution
A hierarchical approach involving child and parent models is used, where child models generate individual datasets and parent models provide latent space data to ensure intercorrelated synthetic datasets, allowing for faster training, flexibility, and reduced resource use.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single model is used to generate multiple intercorrelated datasets, then the model can produce correlated outputs, but the model becomes large and consumes excessive computational resources
Solution Approach 1:
The patent divides a single large model into multiple smaller child models, each responsible for generating a specific dataset. These child models are trained independently on their respective datasets but share a common latent space through parent models, enabling them to generate correlated data without requiring one massive model. This segmentation reduces computational resource consumption while maintaining correlation accuracy.
2Reliability
If a single model is used to generate multiple intercorrelated datasets, then the model can produce correlated outputs, but the model architecture becomes brittle and requires changes when additional datasets are added
Solution Approach 1:
By segmenting the system into independent child models organized in a hierarchical structure with parent models, the patent enables easy addition of new datasets. New child models can be added to the hierarchy without restructuring the entire model architecture, as each child model operates independently while maintaining correlations through the shared latent space.
Solution Approach 2:
The parent models serve as universal components that manage the latent space for multiple child models. This universal structure allows the system to handle various types of datasets and correlations through the same architectural framework, enhancing adaptability when new data sources are introduced.
3Productivity
If conventional separate models are used for each dataset, then training is efficient, but the generated datasets lack proper correlation
Solution Approach 1:
The patent introduces parent models as intermediary components that connect child models through a shared latent space. The parent models generate latent representations that are passed to child models, ensuring that the generated datasets maintain proper correlations while allowing each child model to be trained efficiently on its specific dataset independently.
Data Source
AI summary
Systems and methods for generating synthetic intercorrelated data are disclosed. For example, a system may include at least one memory storing instructions and one or more processors configured to execute the instructions to perform operations. The operations may include training a parent model by iteratively performing steps. The steps may include generating, using the parent model, first latent-space data and second latent-space data. The steps may include generating, using a first child model, first synthetic data based on the first latent-space data, and generating, using a second child model, second synthetic data based on the second latent-space data. The steps may include comparing the first synthetic data and second synthetic data to training data. The steps may include adjusting a parameter of the parent model based on the comparison or terminating training of the parent model based on the comparison