4 ms·
Knowledge graphs where created to solve the problem of making natural,free flowing text machine processable. We now have a technology that completely understand
by itkovian_ 2y ago
Knowledge graphs where created to solve the problem of making natural,free flowing text machine processable. We now have a technology that completely understands natural free flowing text and can extract meaning. Why would going back to structure help when that structure can never be as rich as just text. I get it if the kb has new information, that's not what I'm saying.
- esafak 2y agoBut RAG without graphs just relies on similarity search, which isn't very smart.
- visarga 2y ago> Why would going back to structure help When your corpus is large it is useful to split it up and hierarchically combine. In their place I would do both bottom-up and top-down summarization passes, so information can percolate from a leaf to the root and from the root to a different leaf. Global context can illuminate local summaries, for example think of the twist in a novel, it sheds new light on everything.
- itkovian_ 2y agoThat's not what a kb is
- gkorland 2y ago[dead]
- th0ma5 2y ago> We now have a technology that completely understands natural free flowing text and can extract meaning. Actually we don't. I know it certainly feels like LLMs do this but no one would dare stake their life on their output if they know how they work. Still useful!