This invention outlines a method to create a specialized knowledge graph for a particular use. It begins by extracting facts and relationships from a user's own data, such as conversations, to form a basic network. This initial network is then expanded by finding related information in a larger, existing knowledge base, specifically by identifying disconnected parts of the user's network and linking them to the broader knowledge.
Why it matters: Filed just as Large Language Models were becoming highly capable at entity and relation extraction. The methods described for processing user data and expanding knowledge graphs could now be significantly enhanced and automated by modern LLM capabilities, making the implementation more robust and efficient.
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