Turn abandoned patents
into your own filing
into your own filing
This describes a method to create a new map of relationships between specific entities, like people or businesses, found within a large collection of text documents. It starts with an existing map showing how these documents are related to each other. To decide if two entities should be connected in the new map, the system checks if the documents mentioning those entities are linked in the original document map, using a specific sampling technique. This process is specifically for large sets of unstructured text documents, where entities are often people or businesses.
Why it matters: Filed before the widespread adoption of powerful Large Language Models. The core steps of identifying document relationships and extracting entities from unstructured text, which were challenging in 2017, are now significantly more accurate and efficient with modern NLP and LLM capabilities, making the initial graph construction much more robust.
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
