Memory grounded conversational reasoning and question answering for assistant systems

This invention describes an assistant system that answers user questions by using a personalized "memory graph." When a user asks something, the system identifies key topics, then searches through a graph of the user's past experiences and facts. Using machine learning, it selects the most relevant memories and information from this graph to generate a tailored response. The claims narrow this to selecting candidate nodes specifically from the user's episodic memories.

Why it matters: Filed before the widespread availability and capability of large language models (LLMs). These LLMs now make the "machine-learning models" for memory selection and response generation significantly more powerful and easier to implement than in 2019.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedHeavy software system, machine learning, data structures

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