Use of zettelkasten for agent memory
zettelkasten may be useful for agent idea retrieval and agent idea relation representation for agi memory
The benefits of using zettelkasten for agent memory are that:
- it is human interprable
- facilitates discovery of relevant ideas
- allows for natural language description of ideas
- links between ideas can have meaning (semantic linking)
- doesn’t enforce rigid structure
Problems which need to be solved:
- how for agent to determine when to store an idea
- how to refactor ideas, prune and make connections to allow better navigation
- delete ideas and combine ideas
- find relevant ideas to link to
- determine when to stop searching through memory
If a zettelkasten is employed:
- for each idea/memory, an agent should store semantic links to other ideas
- these can be added or removed over time as new connections/relevant ideas are discovered
- allows for more efficiently finding relevant ideas when solving problem/carrying out a task
- these can be added or removed over time as new connections/relevant ideas are discovered
- more thoughts on this topic can be found in learning as adding new functions to a concept network and AI agent memory optimisation design
If you found this interesting, have feedback or are working on something related, let’s chat: twitter (@0xdist) or schedule a 20 min call
Distbit
Interested in econ, cryptoecon, agents, finance, epistemology, liberty.