Conclusions


It is possible to find representations of events, episodes and contexts which satisfy all the requirements for a system-level model of episodic memory.

This has involved developing several new concepts:

1. A representation of events based on neuroanatomy

2. A representation of episodes which satisfy chunk size constraints and with characterizing keys, which are used for retrieval and for formation of contexts

3. A hierarchical representation of contexts and a process by which they are applied to a given situation to solve problems.

4. The concepts of context and result are different from previous AI concepts.
All of these mechanisms and representations are designed to work on our system-level model of the brain, which is a set of interconnected modules with associative memories.