Reasonable language model learning for text generation from a knowledge graph

This invention describes a computer system that learns to generate text by using structured information from a knowledge graph. It works by analyzing facts within the knowledge graph and other data sources to create specialized training examples. These examples link specific facts to descriptive labels, which are then used to teach language models how to produce relevant and coherent text. The training data is specifically generated by identifying keywords from the knowledge graph's facts and finding matching labels in other data sources.

Why it matters: Filed before the widespread adoption of powerful large language models. The method of training language models using knowledge graphs is now significantly more effective with advanced LLM architectures and fine-tuning techniques.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedHeavy software system, AI/ML, data engineering

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