IBM Technology
Ready to become a certified watsonx AI Assistant Engineer? Register now and use code IBMTechYT20 for 20% off of your exam → https://ibm.biz/BdngMV
🚀 Try GraphRAG now! Access the code here → https://ibm.biz/BdngaC
Learn more about GraphRAG here → https://ibm.biz/BdngM9
🤖 Can AI turn text into structured knowledge? Discover how GraphRAG leverages knowledge graphs, graph databases, and Cypher queries to transform unstructured data into actionable insights. See how LLMs enable intelligent retrieval and automation, reshaping workflows across industries. 🚀
AI news moves fast. Sign up for a monthly newsletter for AI updates from IBM → https://ibm.biz/BdngMU
#knowledgegraph #cypher #ai
Source
Great stuff… I’m in the process of building an app and I was looking for ways to create a graph db from unorganized content 🎉
Thank you for this information. Amazing
Great ❤❤❤ work
Thanks for sharing
IBM makes the highest number of AI videos but has one of the lowest rated AI Models. Strange world.
Nice. Maybe this is a wrong question, how would I connect/build this with IBM’s Bee agent framework.
Thanks for code.
100% must try.
I much prefer using a Graph store than a vector because you it skips the process of creating embeddings which are model specific, meaning that any language model can use data in a Graph, so you don't have to redo your embeddings everytime you change your model.
M loving these videos. It helps build my knowledge pool
Nice
Great! Something new that I wasn't familiar before
how to automate and generate "allowed_nodes" and "allowed_relationships"? doing this manually kinda defeats its purpose imho …
The best!!! Amazing!
Use dark mode folks!
I stopped watching a little after 2:00
Won't using LLM to generate the graph from the input data involve significant pre-processing costs?
What, if I have to, load some 800 pages of data into a graph??
Wonderful! any version using the docker desktop or the procedure is the same?
Thank for sharing
Great video! Thanks for sharing
Need to include diagrams and not just talk about it. Its boring.
How do i use this flow with other languages? Like French for example. I tried to use Mistral model but somehow, it doesn't work.
Why not just use old good SQL on relational DB instead of all this overhead? Before they prompted all those LLMs, I would already acoumplisged whole job with simple SQL