Synthetic data generation in computer-based reasoning systems

This invention describes a computer method for creating fake training data for AI systems. It works by picking a few key examples from existing data, then generating new data points based on those examples. A crucial step is checking if these newly generated data points are too similar to any existing data, often using a distance measurement, to ensure the new data adds genuine variety.

Why it matters: Since 2022, generative AI models like LLMs have made synthetic data generation much more powerful. This invention's focus on ensuring generated data isn't overly similar to existing data could be crucial for controlling these advanced models, preventing overfitting, and enhancing data diversity.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedSoftware logic, machine learning algorithms

3 AI reinvention directions for this patent

See all three

AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.

Reinvent this with AI

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10