Enhancing performance of RAG systems with mathematical methods

This event is reserved for our members, but we still have a few places available for those who would like to discover the club. Don't hesitate to sign up - you'll be put on a waiting list and we'll confirm your place a days before the event.

Matthieu Olekhnovitch, CTO @laive.ai Florian Palmade, CEO @laive.ai

RAG still struggles when facing complex B2B environments, incl. large documents bases.

As a result, AI agents that depend on RAG are not reliable, and a lot of users don't trust them.That's the issue the Laive.ai has decided to tackle : enhancing RAG system to improve the relevance of AI Agents' input data.

The main problem they address is the poor relevance of retrieved chunks when similarity is performed on a huge base of documents, which is common in B2B use cases.

During this workshop, they will present the different techniques of agentic retrieval that they are working on in order to improve reliability, including :

- Different architectures depending on the type of documents and use cases

- More meta data saved when ingesting / indexing documents chunks

- Analysis of the links between documents, from a logical point of view, in order to better understand the structure of the documents base (e.g. using graphs)

- Pre-filtering of embedded data before retrieval- ... and more !

They will also present some use cases and first tests they have conducted.

2510-Laive-BetterRAG

Know what’s Happening Next — before everyone else does.

Get App

Host or Publisher datacraft