Machine learning for joint recognition and assertion regression of elements in text

This invention describes a computer system that can read unstructured text, like a document or an email. It uses a specialized machine learning program to first find important things or people (entities) in the text. Then, using the same program, it figures out what is being said or asserted about those entities. The program works by breaking the text into small pieces, understanding their meaning in context, and then representing the assertions as numerical values in a multi-dimensional space. The claims specifically narrow this to using a contextual language model, like BERT, for this process.

Why it matters: Filed before the widespread adoption and advanced capabilities of large language models (LLMs) became apparent. Modern LLMs can now perform complex contextual understanding, named entity recognition, and assertion extraction with greater accuracy and often with simpler architectures than the specific BERT-based, post-CLM layer approach described in 2020.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedSoftware logic and ML expertise

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