Generation of optimized spoken language understanding model through joint training with integrated knowledge-language module

This invention describes a method to create a highly effective speech understanding system. It works by simultaneously training a language understanding component and a speech recognition component. The language component is further enhanced by integrating factual knowledge from a knowledge graph, allowing the system to leverage both acoustic information and deep language understanding, including factual context, for better interpretation of spoken words.

Why it matters: Filed before the widespread adoption and capabilities of large language models (LLMs) were fully apparent. The significant advancements in LLMs and multimodal AI since 2021 could now provide much more powerful language and knowledge modules, making the integrated training described here more effective and the resulting speech understanding model far more capable.

Status: ExpiredProtection ended and was never renewed
How hard to build: SpecializedAI model training, large datasets, computing infrastructure

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