Vector Databases and Symbolic logic and Large Language Models may hold the secret to episodic memory

In this hypothetical scenario where a vector database stores episodic symbols of memory in time frames and a Large Language Model (LLM) learns from the symbolic graphs, we can theorize about the capabilities of such a machine. The vector database would store episodic memories as symbolic graphs representing events in time frames. Each node in the graph would contain information about the event, such as time, place, emotions, and contextual details. The LLM would be trained to understand and interpret the symbolic graphs stored in the vector database. It would learn the patterns, relationships, and context within the episodic memories to make sense of the events encoded in the graphs. By analyzing the episodic symbols in the vector database, the machine could construct mental representations of past events and scenes. It would be able to generate a coherent and detailed picture of the events based on the stored symbolic information.