This story was originally published on HackerNoon at: https://hackernoon.com/my-dandd-campaign-notes-accidentally-matched-googles-new-rag-spec.
If your data already has structure that only humans read, the gap between "looks like a graph" and "is a graph your retrieval code uses" is worth checking.
Check more stories related to machine-learning at: https://hackernoon.com/c/machine-learning.
You can also check exclusive content about #rag, #retrieval-augmented-generation, #artificial-intelligence, #machine-learning, #google-cloud, #vector-search, #knowledge-graph, #hackernoon-top-story, and more.
This story was written by: @jamiesmith. Learn more about this writer by checking @jamiesmith's about page,
and for more stories, please visit hackernoon.com.
I built a budget RAG system over my D&D group's Obsidian vault, hit the same retrieval failures everyone hits, tried giving the model an agent loop to fix them, then tested Google's new Open Knowledge Format graph structure against the failures neither fixed. One of them worked. Most of them didn't. Here's the honest version.