Associative memory system for LLMs that learns during inference [P]

About this article

I've been working on MDA (Modular Dynamic Architecture), an online associative memory system for LLMs. Here's what I learned building it. The problem I was trying to solve RAG can't learn mid-conversation. If you introduce a new fact after indexing, it's invisible to retrieval. I wanted a system that could learn during inference without retraining. How MDA works Every concept becomes an Entity with a 256-dim identity vector. Entities are connected through a sparse synapse graph. New knowledge...