Synthetic Data Generation via Logical Graph Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge is to generate accurate and representative training data for AI models without relying on real-world data, which is increasingly restricted by regulations and may contain hidden deficiencies.
Innovation Solution
A method and system that guide users to create a logical graph representing how real data is generated, using a GUI, and simulate scenarios with an AI model to produce synthetic data, allowing for iterative refinement and inclusion of nuanced steps and rewards to capture real-world patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real data is collected and used for training AI models, then the training data accuracy and representativeness are improved, but regulatory compliance difficulties and data availability limitations increase
Solution Approach 1:
The patent creates synthetic copies of real-world data through simulation rather than using actual real data. The system generates artificial training data that replicates the statistical properties, relationships, and patterns of real data while avoiding all regulatory and availability issues associated with collecting and using genuine real-world data.
Solution Approach 2:
The patent introduces simulation technology as an intermediary between the need for accurate training data and the limitations of real data availability. Rather than directly using real data or working without accurate data, the simulation acts as a mediator that produces synthetic data with the desired accuracy properties without the drawbacks of real data collection.
2Quantity of substance
If real data is collected through extensive processes, then data completeness is improved, but process complexity and time consumption increase
Solution Approach 1:
The patent performs preliminary actions by pre-defining the simulation environment, rules, and parameters before data generation begins. Rather than collecting data through complex ongoing processes, the system is configured in advance to automatically generate complete and representative synthetic data through the simulation execution.
3Reliability
If real data is used for training, then data representativeness is improved, but hidden deficiencies and regulatory risks increase
Solution Approach 1:
The patent converts the potential harm of regulatory non-compliance and hidden deficiencies in real data into a benefit by using synthetic data generation. The simulation approach eliminates regulatory risks entirely while maintaining or improving data representativeness, turning what would be a harmful situation into a beneficial solution.
4Measurement precision
If complex situations are simulated in full detail, then data accuracy is improved, but computational complexity and simulation time increase
Solution Approach 1:
The patent applies partial action by implementing progressive refinement of simulation detail. Rather than simulating every possible detail at maximum complexity from the start, the system begins with essential elements and adds detail progressively where needed, achieving sufficient accuracy without unnecessary computational overhead.
Data Source
AI summary
Using a graphical user interface (GUI), a user is guided to generate a logical graph that represents how real data is generated in a situation. The situation is simulated using an artificial intelligence (AI) model a number of times by having the AI model choose paths through the logical graph to generate synthetic data that is representative of the real data. The AI model may also use a reward table that the user is guided to generate via the GUI to execute the simulation, wherein the reward table incentivizes certain paths of the logical graph, and agents of the AI model are trained using the values of this reward table. This synthetic data generated via the AI model can be used for training other AI models.