Neuro-Symbolic AI Summer School – Dr. Aasman’s Presentation

“Context Graphs as Neuro-Symbolic Memory for Reliable AI Agents” – Dr. Jans Aasman’s (CEO, Franz Inc.) presentation at the Neuro-Symbolic AI Summer School, August 6

At the 2026 Neuro-Symbolic AI Summer School, Dr. Jans Aasman, CEO of Franz Inc., presented “Context Graphs as Neuro-Symbolic Memory for Reliable AI Agents.”

The presentation focused on a central challenge for agentic AI: LLMs are powerful at interpreting language and recognizing patterns, but they need structured knowledge, reasoning, and persistent memory to operate reliably in enterprise environments.

This is where Neuro-Symbolic AI becomes critical.

Neural technologies such as LLMs and embeddings provide flexibility and language understanding. Symbolic technologies such as Knowledge Graphs, rules, semantics, and reasoning provide structure, constraints, provenance, and explainability.

A Context Graph brings these capabilities together by extending the Knowledge Graph with operational state, decision history, temporal relationships, objectives, policies, and agent memory.

The result is an architecture where AI agents can:

understand → reason → act → observe → remember

AllegroGraph provides this Neuro-Symbolic foundation by combining Knowledge Graphs, vector search, semantic reasoning, Prolog rules, and GraphTalker’s agentic natural-language interface.

The goal is not simply better answers from AI, but more reliable agents that can reason over trusted knowledge, preserve context over time, and explain how they reached a conclusion.