In this episode, join Angelina and Mehdi, for a discussion about another new technique for advanced RAG - RAPTOR. In essence, it combines hierarchical clustering technique with information retrieval through RAG.

Who's Angelina:   / meetangelina  
Who's Mehdi:   / mehdiallahyari  


00:00 "Retrieval augmented language models can better adapt to changes in world state and incorporate long tail knowledge"
01:11 Why RAPTOR?
01:26 Brief RAG revisit
02:02 Challenges for vanilla RAG
02:10 Comparison questions and "how" questions
03:46 The Cinderella question
04:46 How does RAPTOR work?
06:29 Visualizing the model architecture
09:07 How's it related to the Sub-Doc method?
11:42 Summary of summaries
14:37 The analogy of a book (again!)
16:33 Ladder of Abstraction
18:04 Is it worth it to put it in production?
22:24 When is it best suited for production use?
23:57 Bonus: Two tree retrieval methods
24:05 Bonus: Tree Traversal Retrieval
27:57 Bonus: Collapsed Tree Retrieval

🦄 Any specific contents you wish to learn from us? Sign up here: https://noteforms.com/forms/twosetai-...

🖼️ Blogpost for today:
Ladder of Abstraction: https://open.substack.com/pub/mlnotes...

🗞️Paper: RAPTOR: RECURSIVE ABSTRACTIVE PROCESSING FOR TREE-ORGANIZED RETRIEVAL: https://arxiv.org/pdf/2401.18059.pdf

🔨 Implementation: Coming next week!

📬 Don't miss out on the latest updates - Subscribe to our newsletter: https://mlnotes.substack.com/

📚 If you'd like to learn more about RAG systems, check out our book on the RAG system: https://angelinamagr.gumroad.com/

🕴️ Our consulting firm: We help companies that don't want to miss the boat of the current wave of AI advancement by integrating these solutions into their business operations and products. https://www.transformaistudio.com/

Stay tuned for more content! 🎥 Thanks you for watching! 🙌