Using Elasticsearch for Multimodal Search

Multimodal search, where images and text are combined to form a powerful search, is a rapidly emerging trend.

Retail segments, such as fashion and home design, are one particular driver of multimodal search because they rely heavily on visual search since style is often difficult to describe using text.

However, in addition to visual search, text search is still a required part of the solution because product information, such as item description, item title, category, and brand are generally used to filter the results that are returned as part of the visual search. Thus what is needed is a solution that allows for multimodal search.

In this presentation, you will learn about an Elasticsearch plugin that:
· Integrates seamlessly with native Elasticsearch text search to provide multimodal search
· Uses the native Elasticsearch dense_vector field to perform approximate nearest neighbor vector similarity search
· Requires no reindexing of documents to support vector search
· Scales to billion-scale vector similarity search

Timestamps:
00:00 Introduction
02:01 What is Multimodal Search?
03:24 What are Embeddings?
05:43 Who is Using Embeddings and Dense Vectors?
07:08 Dense Vectors in Elasticsearch
09:25 Miltimodal Elasticsearch
10:18 Limitations of Multimodal Elasticsearch
11:46 GSI's Hardware Approach & Demo
18:50 Architecture - GSI Elasticsearch Engine
23:54 Elastisearch Benhmarks
26:57 Conclusion & Final Remarks
28:57 Q&A

Speakers: George Williams and Tal Refaeli

#elasticsearch #multimodal #search #multimodalai #ai #embedding #densevector #gsi

[raghupathym]

Thank you, How to upload images in elasticsearch and visualize it in kibana?