Neural Search - From Prototype to Production with Jina

Neural Search - From Prototype to Production with Jina

by Bo Wang, Cristian Mitroi, Feng Wang, Shubham Saboo, Susana Guzmán
Released October 2022
Publisher(s): Packt Publishing
ISBN: 9781801816823

Read it now on the O’Reilly learning platform with a 10-day free trial.

O’Reilly members get unlimited access to books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.

Start your free trial

Book description

Implement neural search systems on the cloud by leveraging Jina design patterns

Key Features

  • Identify the different search techniques and discover applications of neural search
  • Gain a solid understanding of vector representation and apply your knowledge in neural search
  • Unlock deeper levels of knowledge of Jina for neural search

Book Description

Search is a big and ever-growing part of the tech ecosystem. Traditional search, however, has limitations that are hard to overcome because of the way it is designed. Neural search is a novel approach that uses the power of machine learning to retrieve information using vector embeddings as first-class citizens, opening up new possibilities of improving the results obtained through traditional search.

Although neural search is a powerful tool, it is new and finetuning it can be tedious as it requires you to understand the several components on which it relies. Jina fills this gap by providing an infrastructure that reduces the time and complexity involved in creating deep learning–powered search engines. This book will enable you to learn the fundamentals of neural networks for neural search, its strengths and weaknesses, as well as how to use Jina to build a search engine. With the help of step-by-step explanations, practical examples, and self-assessment questions, you'll become well-versed with the basics of neural search and core Jina concepts, and learn to apply this knowledge to build your own search engine.

By the end of this deep learning book, you'll be able to make the most of Jina's neural search design patterns to build an end-to-end search solution for any modality.

What you will learn

  • Understand how neural search and legacy search work
  • Grasp the machine learning and math fundamentals needed for neural search
  • Get to grips with the foundation of vector representation
  • Explore the basic components of Jina
  • Analyze search systems with different modalities
  • Uncover the capabilities of Jina with the help of practical examples

Who this book is for

If you are a machine learning, deep learning, or artificial intelligence engineer interested in building a search system of any kind (text, QA, image, audio, PDF, 3D models, or others) using modern software architecture, this book is for you. This book is perfect for Python engineers who are interested in building a search system of any kind using state-of-the-art deep learning techniques.

Table of contents

  1. Neural Search – From Prototype to Production with Jina
  2. Contributors
  3. About the authors
  4. About the reviewer
  5. Preface
    1. Who this book is for
    2. What this book covers
    3. To get the most out of this book
    4. Download the example code files
    5. Download the color images
    6. Conventions used
    7. Get in touch
    8. Share Your Thoughts
  6. Part 1: Introduction to Neural Search Fundamentals
  7. Chapter 1: Neural Networks for Neural Search
    1. Technical requirements
    2. Legacy search versus neural search
      1. Exploring various data types and search scenarios
      2. How does the traditional search system work?
      3. Pros and cons of the traditional search system
    3. Machine learning for search
      1. Understanding machine learning and artificial intelligence
      2. Machine learning and learning-to-rank
    4. Practical applications powered by neural search
    5. New terms learned in this chapter
    6. Summary
  8. Chapter 2: Introducing Foundations of Vector Representation
    1. Technical requirements
    2. Introducing vectors in ML
      1. Using vectors to represent data
    3. Measuring similarity between two vectors
      1. Metrics beyond cosine similarity
    4. Local and distributed representations
      1. Local vector representation
      2. Distributed vector representation
    5. Summary
    6. Further reading
  9. Chapter 3: System Design and Engineering Challenges
    1. Technical requirements
    2. Introducing indexing and querying
      1. Indexing
      2. Querying
    3. Evaluating a neural search system
    4. Engineering challenges of building a neural search system
    5. Summary
  10. Part 2: Introduction to Jina Fundamentals
  11. Chapter 4: Learning Jina’s Basics
    1. Technical requirements
    2. Exploring Jina
    3. Documents
      1. Document attributes
    4. DocumentArray
      1. Constructing a DocumentArray
    5. Executors
      1. Creating an Executor
    6. Flow
      1. Creating a Flow
      2. Adding Executors to a Flow
    7. Summary
  12. Chapter 5: Multiple Search Modalities
    1. Technical requirements
    2. Introducing multimodal documents
      1. Text document
      2. Image document
      3. Audio document
      4. Multimodal document
    3. How to encode multimodal documents
      1. Encoding text documents
      2. Encoding image documents
      3. Encoding audio documents
    4. Cross-modal and multimodal searches
      1. Cross-modal search
      2. Multimodal search
    5. Summary
  13. Part 3: How to Use Jina for Neural Search
  14. Chapter 6: Building Practical Examples with Jina
    1. Technical requirements
    2. Getting started with the Q/A chatbot
      1. Navigating through the code
    3. Understanding fashion image search
      1. Navigating through the code
    4. Working with multimodal search
      1. Navigating through the code
    5. Summary
  15. Chapter 7: Exploring Advanced Use Cases of Jina
    1. Technical requirements
    2. Introducing multi-level granularity
      1. Navigating through the code
      2. app.py
      3. index.yml
      4. query.yml
      5. Installing and running the example
    3. Cross-modal search with images with text
      1. app.py
      2. flow-index.yml
      3. query.yml
      4. Installing and running the example
    4. Concurrent querying and indexing data
      1. app.py
      2. flow.yml
      3. Installing and running the example
    5. Summary
  16. Index
    1. Why subscribe?
  17. Other Books You May Enjoy
    1. Packt is searching for authors like you
    2. Share Your Thoughts

Product information

  • Title: Neural Search - From Prototype to Production with Jina
  • Author(s): Bo Wang, Cristian Mitroi, Feng Wang, Shubham Saboo, Susana Guzmán
  • Release date: October 2022
  • Publisher(s): Packt Publishing
  • ISBN: 9781801816823