Replying to 360e9b8e...

Today's RAG experiments using #SPARQL, #DBpedia, and our #VirtuosoRDBMS + #ChatGPT based Personal Assistant, leverages native text indexing.

Initial prompt:

Where did Euler grow up?

Initially, I had to get the agent to remember the rules for invoking SPARQL-FED. #HowTo #AI
36
360e9b8e... 2y ago

Then I posed a more challenging (vaguer) question:

What is Euler credited for popularizing?

Initially, that threw the #SPARQL generator off. Then I asked #ChatGPT to hone into the dbo:abstract attribute which comprises indexed text, which worked!

#RAG #HowTo #KnowledgeGraph

Reply to this note

Please Login to reply.

Discussion

36
360e9b8e... 2y ago

Here's a link to the transcript for the aforementioned #SPARQL-based RAG session that demonstrates loose coupling with the #DBpedia #KnowledgeGraph.

https://netid-qa.openlinksw.com:8443/chat/?chat_id=697862bdf680558773de22ba7014da03 -- via

nostr:npub15ntm6s6ytnamu9959rqjydzz699xss6dwuy7kawxxfr2k8nr7rzq8fuxn0 Personal Assistant

#RAG #LinkedData #SemanticWeb