agent memory
|
+--------------+--------------+
| |
deductive state episodic memory
| |
facts / rules / time fuzzy context
provenance semantic retrieval
retractions source text
I accidentally turned LLM memory into program analysis
101–104 of 104 posts
Re: I accidentally turned LLM memory into program analysis
#102I reached a similar conclusion: LLMs should only really sit at the terminals of request fulfilment. 1. User request understanding: natural language -> a more rigorous representation, in my case Datalog. 2. Result interpretation: facts and derived facts -> natural language. Between those terminals, the work should be mechanical reasoning over some ontology or formal knowledge structure. That connects to another princi…
When i hear “weathering” i think of something slowly eroding away
Weather is fun choice since it’s a contronym. It has opposing definitions: to endure and to erode.
“He weathered the storm” vs “the stone weathered into an arch”.
Re: I accidentally turned LLM memory into program analysis
#103I’ve been playing with this too, the OP is rediscovering the whole field of AI before LLMs. Prolog, temporal logic …
The next one is ontology and semantic web, for which the “trying to stem verbs and nouns” is a crude v0.1 attempt.
Re: I accidentally turned LLM memory into program analysis
#104[flagged]