Class-Conditioned Tabular Data Generation With Fidelity Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing generative models struggle to preserve class-conditional feature distributions and scale effectively across diverse class labels in structured, tabular datasets, leading to generic outputs lacking diversity or fine-grained fidelity.
Innovation Solution
A recursive, stage-wise generative framework that injects structured noise into training datasets and applies class-conditioned, multi-output decision tree ensembles to generate high-fidelity, label-consistent synthetic data, with a fidelity monitor evaluating statistical divergence to trigger selective retraining.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional generative approaches are used, then data generation can be performed, but class-conditional feature distributions are not preserved and outputs lack diversity and fidelity
Solution Approach 1:
The patent segments the generative process by introducing structured noise at specific stages and using class-specific noise injection. This segmentation allows different parts of the data generation process to handle different aspects: the base model generates general structures while class-specific noise injection preserves conditional distributions, thereby resolving the contradiction between fidelity and adaptability.
Solution Approach 2:
The patent applies local quality by injecting noise with characteristics specific to each class rather than using uniform noise across all classes. This allows the synthetic data to maintain high fidelity for each specific class condition while preserving the diversity needed across different class labels, thus resolving the contradiction between manufacturing precision and adaptability.
2Productivity
If generic generative models are used, then scalability across diverse class labels is achieved, but outputs lack fine-grained fidelity and diversity
Solution Approach 1:
The patent changes parameters dynamically by adjusting noise characteristics based on class labels. The system maintains scalability through a unified framework while achieving fine-grained fidelity by modifying noise parameters specific to each class, thus resolving the contradiction between productivity and manufacturing precision.
Solution Approach 2:
The patent introduces dynamics by making the noise injection process adaptive rather than static. The system dynamically selects and applies noise characteristics based on the target class label, enabling it to scale across diverse classes while maintaining fine-grained fidelity for each specific class condition.
3Manufacturing precision
If class-specific noise injection is applied, then diversity and fidelity of synthetic data are improved, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-defining noise characteristics and class-specific parameters before the actual data generation process. This preparation work is done once and then reused across multiple generation tasks, reducing the computational complexity during actual synthesis while maintaining high diversity and fidelity through the pre-configured class-specific noise injection.
Data Source
AI summary
An example operation may include at least one of injecting, by a noise injection module, structured noise into input data, selecting, by at least one processor communicatively coupled to a memory on a host platform, a class-specific model from a set of trained tree-based generators, evaluating, by a fidelity monitor, synthetic tabular data, wherein the fidelity monitor transmits a retraining signal to an AI development system when the synthetic tabular data deviates from expected distributional patterns, identifying, by the AI development system, a target model based on the retraining signal, retraining, by the AI development system, the target model based on the retraining signal, transmitting, by the AI development system, a retrained model to an AI production system, replacing, by the AI production system, a deployed model with the retrained model, receiving, by the AI production system, a query from a computing device, and responding, by the AI production system, to the query using the retrained model.