Relevance Feedback
Properties
tags
cscs/databases
created
02.03.2026, 00:00
modified
26.07.2026, 13:43
published
Empty
sources
Empty
topics
Relevance Feedback, Vector Search, Information Retrieval
authors
Empty
ai-assisted
Yes
- Relevance feedback - Wikipedia
- Relevance feedback is an iterative process that refines search results based on user interactions. After a system returns initial results, the user marks specific documents as relevant or irrelevant. The system then updates the query and searches again to improve precision and recall.
- Three types of feedback: explicit feedback, implicit feedback, and blind or “pseudo” feedback.
# Resources
- Relevance Feedback in Qdrant
- 2026-02-19
- Relevance Feedback in Informational Retrieval
- 2025-03-26
# Discussions With AI
- Claude Discussion - Relevance feedback & Wormhole Vectors
- Walk-through of how the Qdrant relevance feedback formula replaces cosine similarity on the second pass