Intelligent control with hierarchical stacked neural networks

This invention describes a way to spot unusual or "aberrant" messages by analyzing their grammar. It works by taking a natural language message, identifying its words, and comparing them to a set of expected words with known meanings. Then, it checks how well the message's grammatical structure matches what's expected for that type of message, specifically for messages with a high level of complexity, flagging any significant deviations as "cognitive noise."

Why it matters: Filed before the widespread availability of powerful large language models (LLMs), the complex semantic and grammatical analysis described is now significantly more feasible and accurate using current LLM capabilities. These models can readily perform the detailed linguistic pattern recognition and deviation detection that was much harder to achieve in 2019.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedNatural language processing, neural networks, heavy software

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