the backbone of the entire application is a huge knowledge base graph that semantically connects your information based on topics, keywords, participants, different relationships, etc
and then the search would be a GraphRAG with other techniques as well.
of course, LLMs are used to vectorize strings, extract information, etc, but it's more complicated than an LLM integration.
I really like the idea.
thank you.
do you also experience issues like these in your daily work?
this really looked like an llm integration
unfortunately, it's way more complicated.
the backbone of the entire application is a huge knowledge base graph that semantically connects your information based on topics, keywords, participants, different relationships, etc
and then the search would be a GraphRAG with other techniques as well.
of course, LLMs are used to vectorize strings, extract information, etc, but it's more complicated than an LLM integration.
ah makes sense. thanks for clarifying!