You know a topic is complex when someone writes a second article to explain the first, but this article on vector databases and RAG does a great job of it!

The LinkedIn article titled Vector Databases and RAG (written by Simson Garfinkel and available here) was recommended to me last week by Craig Ball, who called it “one of the most lucid explanations of high dimensional search I’ve ever read”. That certainly got my attention!

Of course, last week I was busy at the UF-Law E-Discovery Conference, so I just now got around to reading it today. Garfinkel’s article is designed to provide context and clarity to the blog post You Don’t Need a Vector Database, written by Dr. Yucheng Low, co-founder of XetHub. In his article, Garfinkel explains what a vector database is, explains why Dr. Low says you don’t need one – for Retrieval Augmented Generation (RAG), that is – and provides context to Dr. Low’s answer.

A couple of weeks ago, I discussed RAG, what it is and how it works (at a high-level). But this diagram from Dr. Low’s article (created by the Scriv.ai team) does a great job of illustrating it:

However, to make sense of Dr. Low’s article, Garfinkel discusses why vectors of numbers have become such an important tool for search systems, and why that popularity can translate into unnecessary expense for organizations that deploy vector-based systems. That discussion includes an illustration of vectors of numbers, bags of words, word embeddings and vector databases.

The background that Garfinkel provides makes it easier to understand Dr. Low’s article, where he compares three text retrieval approaches and concludes that a hybrid approach of combining traditional text retrieval algorithms with vector technology can produce better search results for RAG than simply using a vector database.

Make sense? No? Haha, I didn’t think it would by simply reading this post. 😉

But it will hopefully make a lot more sense if you read Garfinkel’s article and then Dr. Low’s post. Garfinkel’s article on vector databases and RAG does a great job of explaining the building blocks that go into today’s AI. Check out both articles via the links above! What’s our vector, Victor? 😀

So, what do you think? Do you agree with the findings regarding vector databases and RAG? Please share any comments you might have or if you’d like to know more about a particular topic.

Image created using Bing Image Creator Powered by DALL-E, using the term “robot wiping down a computer monitor with a rag”.

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.