深度学习与搜索：革新信息检索

需积分: 21 190 浏览量更新于2024-07-17 1 收藏 8.21MB PDF 举报

"Deep Learning for Search.pdf" 本书"Deep Learning for Search"由Tommaso Teofili撰写，2019年由Manning出版社发行，是一部全英文的深度学习与搜索技术结合的专业著作。作者在自然语言处理（NLP）领域拥有深厚的背景，他对计算机理解和解析大量文本的能力感到着迷。同时，他在实际工作中接触并应用了开源搜索引擎Lucene和Solr，经历了从调整算法以适应用户查询到手动干预以提升用户体验的过程。在深入学习领域，作者通过Andrew Ng的在线机器学习课程以及Andrej Karpathy和Richard Socher关于深度神经网络的课程自我学习，逐渐积累了丰富的经验，并专注于开放源代码项目中的搜索、NLP和深度学习研究。书中提到，尽管深度学习已经在计算机视觉和NLP方面取得了革命性的突破，但在搜索领域的应用仍有很大的探索空间。虽然我们不能完全依赖深度学习自动配置和优化搜索引擎，但它可以帮助我们改进用户体验，实现传统技术无法做到的功能，并增强现有搜索技术的效果。深度学习为提高搜索引擎的效率开启了新的可能。本书内容可能涵盖深度学习的基础概念，如何应用于搜索引擎的索引、查询匹配和排序算法，以及如何利用深度神经网络解决搜索中的拼写纠错、语义理解等问题。此外，还可能讨论了实践中遇到的挑战，如数据准备、模型训练、性能优化和实时更新等。 "Deep Learning for Search"是面向那些希望将深度学习技术融入搜索解决方案的开发人员和研究人员的一本实用指南，旨在帮助他们提升搜索引擎的准确性和用户体验。通过阅读此书，读者可以了解到深度学习如何改变搜索领域，以及如何在自己的项目中实施这些先进方法。

preface

The field of natural language processing bewitched me as soon as I came to know

about it nearly 10 years ago, while studying for my master’s degree. The promise that

computers could help us understand the (already, even then) vast amount of textual

documents in existence sounded like magic. I still remember how exciting it was to

see my first NLP programs extract even vaguely correct and useful information from a

few text documents.

About the same time, at work, I was asked to do some consulting for a customer on

their new open source search architecture. My colleague, who was an expert in the

field, was busy on another project, so I was given a copy of Lucene in Action,

which I

studied for a couple of weeks; then I was sent out on the consulting job. A couple of

years after I worked on that Lucene/Solr-based project, the new search engine went

live (and, as far as I know, it’s still used). I can’t tell you how many times the search

engine algorithms needed to be adjusted because of this or that query or this or that

fragment of indexed text, but we made it work. I could see users’ queries, and I could

see the data that was there to be retrieved, but a minimal difference in spelling or

omitting a certain word could cause very relevant information to not show up in the

search results. So while I was very proud of my work, I kept wondering how I could

have done better to avoid the many manual interventions the product managers asked

me to perform in order to provide the best possible user experience.

Right after this, I quite by chance found myself involved in machine learning

thanks to Andrew Ng’s first machine learning online class (which originated the Cour-

sera MOOC series). I was so fascinated with the concepts behind the neural networks

shown in the class that I decided to try to implement a small library for neural

Michael McCandless, Erik Hatcher, and Otis Gospodnetic´ (Manning, 2010), http://www.manning.com/

books/lucene-in-action-second-edition.

xvii

acknowledgments

First and foremost I would like to thank my lovely wife Michela for encouraging and

supporting me throughout this long journey: thanks for the love, energy, and dedica-

tion during long days, nights, and weekends of writing!

Thanks go to Giacomo and Mattia for helping me choose the coolest cover illustra-

tion possible and for all the playing and laughs while I was trying to write.

I would like to thank my father for his pride and his confidence in me.

Big thanks go to my friend Federico for his tireless effort in reviewing all the book

material (book, code, images, and so on) and for the enjoyable discussions and ideas

shared. More huge thanks go to my friends and colleagues Antonio, Francesco, and

Simone for their support, laughs, and advice. Thanks also go to my fellow Apache

OpenNLP (http://opennlp.apache.org) friends Suneel, Joern, and Koji for providing

feedback, advice, and ideas that helped shape the book.

I thank Chris Mattmann for writing such an inspiring foreword.

My thanks also go to Frances Lefkowitz, my development editor, for her patience

and guidance throughout the writing process, including our discussions about Steph,

KD, and the Warriors. And I thank the others at Manning who made this book possi-

ble, including publisher Marjan Bace and everyone on the editorial and production

teams who worked behind the scenes. In addition, I thank the technical peer review-

ers led by Ivan Martinovic´ —Abhinav Upadhyay, Al Krinker, Alberto Simões, Álvaro

Falquina, Andrew Wyllie, Antonio Magnaghi, Chris Morgan, Giuliano Bertoti, Greg

Zanotti, Jeroen Benckhuijsen, Krief David, Lucian Enache, Martin Beer, Michael Wall,

Michal Paszkiewicz, Mirko Kämpf, Pauli Sutelainen, Simona Ruso, Srdan Dukic, and

Ursin Stauss—and the forum contributors. On the technical side, thanks go to

Michiel Trimpe, who served as the book’s technical editor; and Karsten Strøbaek, who

served as the book’s technical proofreader.

Finally I’d like to thank the Apache Lucene and Deeplearning4j communities for

providing such excellent tools and for supporting users in a friendly manner.

ABOUT THIS BOOK xix

touched on. Similarly, you aren’t expected to already know about machine or deep

learning. This book will introduce all the required machine learning and deep learn-

ing basics, together with practical tips regarding the application of deep learning to

search engines in production scenarios.

You should be ready to get your hands on the code and extend existing open

source libraries to implement deep learning algorithms to solve search problems.

Roadmap

This book is divided into three parts:

 Part 1 introduces the basic concepts of search, machine learning, and deep

learning. Chapter 1 introduces the rationale for applying deep learning tech-

niques to search problems by touching on problems with respect to most com-

mon approaches to information retrieval. Chapter 2 gives a first example of

how to use a neural network model to improve the effectiveness of a search

engine by generating synonyms from the data.

 Part 2 deals with common search engine tasks that can be better addressed with

the help of deep neural networks. Chapter 3 introduces the use of recurrent

neural networks for generating queries that are alternatives to the ones entered

by users. Chapter 4 addresses the task of providing better suggestions while the

user is typing the query, with the help of deep neural networks. Chapter 5

focuses on ranking models: in particular, how to provide more-relevant search

results using word embeddings. Chapter 6 deals with the use of document

embeddings both in ranking functions and in the context of content recom-

mendation.

 Part 3 takes up more-complex scenarios like deep learning–powered machine

translation and image search. Chapter 7 guides you through giving your search

engine multilanguage capabilities through neural network–based approaches.

Chapter 8 deals with searching a collection of images based on their contents,

empowered by deep learning models. Chapter 9 discusses production-related

topics like fine-tuning deep learning models and dealing with constantly incom-

ing streams of data.

The complexity of the topics and concepts addressed grows over the course of the

book. If you’re new to deep learning, search, or both, I highly recommend reading

chapters 1 and 2 first. Otherwise, feel free to jump around and pick chapters based on

your needs and interests.

About the code

In this book, code snippets are preferred over fully detailed code listings, in order to

provide quick, easy insight into what the code is doing and how. The full source code

can be found on the book’s page on the Manning website: www.manning.com/

books/deep-learning-for-search. The software will also be kept up to date on the

剩余327页未读，继续阅读

OneCrazyStone

粉丝: 5

深度学习与搜索：革新信息检索

搜索和推荐的匹配：sigir18-deep-match.pdf

Deep+Learning+in+Recommendation.pdf

object detection survey deep learning part.pdf

Towards Playing Full MOBA Games withDeep Reinforcement Learning.pdf

字体跳动深度召回Learning A Retrievable Structure for Large.pdf

Adaptive Interaction Modeling via Graph Operations Search.pdf

All in One Bad Weather Removal Using Architectural Search.pdf

Giancarlo Zaccone deep learning with tensorflow pdf

Python Machine Learning. Machine Learning and Deep Learning with Py, scikit

Deep Learning with Applications Using Python

最新资源