NODES AI 2026 - Multi-Agent Shared Graph Memory: Building Collective Knowledge for Agents

This talk features Vaibhava Lakshmi Ravideshik at NODES AI, focusing on "Multi-Agent Shared Graph Memory: Building Collective Knowledge for Agents." The session delves into the collaboration between AI agents using a shared graph-based memory that facilitates persistent, explainable, and cooperative intelligence. The speaker discusses designing an architecture for a shared knowledge graph that allows real-time updates and queries, incorporating elements like conflict resolution, versioning, and provenance tracking. Through live code snippets and examples using Cypher, the talk illustrates how Neo4j can be utilized to enhance agents' collective memory. Attendees will learn to represent agent beliefs, utilize graph queries for decision-making, and manage evolving multi-agent knowledge effectively.