Panini is a way for AI to keep learning new facts without changing its brain by storing them as tiny linked Q&A facts in an external memory.
This paper turns an AI agentβs memory from a flat list of notes into a logic map of events connected by cause-and-time links.