Systems and methods for intelligently curating machine learning training data …

This invention describes a system for improving machine learning-based dialogue services, like chatbots. It works by first analyzing past user questions and commands to create a test set. Then, it gathers new raw training data from various sources. Before using this new data, the system checks its quality by measuring how well it covers different topics and how diverse it is. The system only uses this new data to train its machine learning models if the data meets a minimum quality standard for coverage.

Why it matters: Filed in 2019, before the widespread adoption of large language models (LLMs), this invention's focus on intelligently curating training data for dialogue systems is now even more critical. The exponential growth in LLM size and complexity has amplified the need for efficient, high-quality data acquisition and validation to improve performance and reduce training costs.

Status: Not in forceListed as no longer active.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 models, data pipelines

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