Reranking and query changes | RAG | Nextia Learning

Rerank candidates with a cross-encoder and rewrite queries, and measure what each adds in quality, time and new ways to fail. Where we are: In the last two lessons, we compared keyword, vector and hybrid search, and we added filters for dates and access. But the right passage is often in the list, only too low to reach the top 5. In this lesson, we add a slower second reader that reorders the best candidates, and we let a model rewrite the question. It works like a manager who reads a recruiter's shortlist. Lesson: https://learning.nextia-ai.com/course... Course: RAG: Building AI That Uses Your Data https://learning.nextia-ai.com/course... Chapters: 0:00 Introduction 0:12 A second reader 0:41 The right passage, too low 1:17 Read the two texts together 2:42 What it adds, and what it costs 4:02 Ask for the right words 5:19 Recap Nextia Learning is free: every course and every certificate. Read the full lesson, run the notebook and answer the knowledge checks on the lesson page. En français:    • Видео   #NextiaLearning #LearnAI