Neural Networks and Deep Learning.pdf-资料库

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第1页.png

第1页 / 共293页

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第2页.png

第2页 / 共293页

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第3页.png

第3页 / 共293页

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第4页.png

第4页 / 共293页

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第5页.png

第5页 / 共293页

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第6页.png

第6页 / 共293页

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第7页.png

第7页 / 共293页

32c3f57a-8c3a-4063-9b8b-ae495bab84d7.pdf-第8页.png

第8页 / 共293页

Neural Networks and Deep Learning Michael Nielsen

2016/10/10 Neural networks and deep learning Neural Networks and Deep Learning 欧拉的博客:www.liuhao.me Neural Networks and Deep Learning is a free online book. The book will teach you about: Neural networks, a beautiful biologically-inspired programming paradigm which enables a computer to learn from observational data Deep learning, a powerful set of techniques for learning in neural networks Neural networks and deep learning currently provide the best solutions to many problems in image recognition, speech recognition, and natural language processing. This book will teach you many of the core concepts behind neural networks and deep learning. For more details about the approach taken in the book, see here. Or you can jump directly to Chapter 1 and get started. Neural Networks and Deep earning hat this book is about On the eercises and problems sing neural nets to recognie handwritten digits ow the backpropagation algorithm works mproving the way neural networks learn visual proof that neural nets can compute any function hy are deep neural networks hard to train Deep learning ppendi: s there a siple algorithm for intelligence cknowledgements Frequently sked uestions f you benefit from the book, please make a small donation. suggest , but you can choose the amount. Sponsors Thanks to all the supporters who made the book possible, with especial thanks to avel Dudrenov. Thanks also to all the contributors to the ugfinder all of Fame. Resources ook F Code repository ichael Nielsens project announcement mailing list Deep earning, draft book in preparation, by oshua engio, an oodfellow, and aron Courville http://neuralnetworksanddeeplearning.com/index.html 1/2

2016/10/10 Neural networks and deep learning 欧拉的博客:www.liuhao.me y ichael Nielsen an 1 ncecorpesecesoosceeseneureorsneepernn eernonress supern sorscenseunerreeoonsruononoercnporecensesens ourereeocopsrenuonsoounooseoureneresencoercusepese conce http://neuralnetworksanddeeplearning.com/index.html 2/2

2016/10/10 Neural networks and deep learning at tis ook is aout 欧拉的博客:www.liuhao.me Neural networks are one of the most beautiful programming paradigms ever invented. n the conventional approach to programming, we tell the computer what to do, breaking big problems up into many small, precisely defined tasks that the computer can easily perform. y contrast, in a neural network we dont tell the computer how to solve our problem. nstead, it learns from observational data, figuring out its own solution to the problem at hand. utomatically learning from data sounds promising. owever, until we didnt know how to train neural networks to surpass more traditional approaches, ecept for a few specialied problems. hat changed in was the discovery of techniques for learning in so- called deep neural networks. These techniques are now known as deep learning. Theyve been developed further, and today deep neural networks and deep learning achieve outstanding performance on many important problems in computer vision, speech recognition, and natural language processing. Theyre being deployed on a large scale by companies such as oogle, icrosoft, and Facebook. The purpose of this book is to help you master the core concepts of neural networks, including modern techniques for deep learning. fter working through the book you will have written code that uses neural networks and deep learning to solve comple pattern recognition problems. nd you will have a foundation to use neural networks and deep learning to attack problems of your own devising. prinipleoriented approa One conviction underlying the book is that its better to obtain a solid understanding of the core principles of neural networks and deep learning, rather than a hay understanding of a long laundry list of ideas. f youve understood the core ideas well, you can rapidly understand other new material. n programming language terms, think of it as mastering the core synta, libraries and data structures of a new language. ou may still only know a tiny Neural Networks and Deep earning hat this book is about On the eercises and problems sing neural nets to recognie handwritten digits ow the backpropagation algorithm works mproving the way neural networks learn visual proof that neural nets can compute any function hy are deep neural networks hard to train Deep learning ppendi: s there a siple algorithm for intelligence cknowledgements Frequently sked uestions f you benefit from the book, please make a small donation. suggest , but you can choose the amount. Sponsors Thanks to all the supporters who made the book possible, with especial thanks to avel Dudrenov. Thanks also to all the contributors to the ugfinder all of Fame. Resources ook F Code repository ichael Nielsens project announcement mailing list Deep earning, draft book in preparation, by oshua engio, an oodfellow, and aron Courville http://neuralnetworksanddeeplearning.com/aout.html 1/

2016/10/10 Neural networks and deep learning fraction of the total language - many languages have enormous standard libraries - but new libraries and data structures can be understood quickly and easily. 欧拉的博客:www.liuhao.me y ichael Nielsen an 1 This means the book is emphatically not a tutorial in how to use some particular neural network library. f you mostly want to learn your way around a library, dont read this book Find the library you wish to learn, and work through the tutorials and documentation. ut be warned. hile this has an immediate problem-solving payoff, if you want to understand whats really going on in neural networks, if you want insights that will still be relevant years from now, then its not enough just to learn some hot library. ou need to understand the durable, lasting insights underlying how neural networks work. Technologies come and technologies go, but insight is forever. andson approa ell learn the core principles behind neural networks and deep learning by attacking a concrete problem: the problem of teaching a computer to recognie handwritten digits. This problem is etremely difficult to solve using the conventional approach to programming. nd yet, as well see, it can be solved pretty well using a simple neural network, with just a few tens of lines of code, and no special libraries. hats more, well improve the program through many iterations, gradually incorporating more and more of the core ideas about neural networks and deep learning. This hands-on approach means that youll need some programming eperience to read the book. ut you dont need to be a professional programmer. ve written the code in ython version ., which, even if you dont program in ython, should be easy to understand with just a little effort. Through the course of the book we will develop a little neural network library, which you can use to eperiment and to build understanding. ll the code is available for download here. Once youve finished the book, or as you read it, you can easily pick up one of the more feature-complete neural network libraries intended for use in production. On a related note, the mathematical requirements to read the book are modest. There is some mathematics in most chapters, but its http://neuralnetworksanddeeplearning.com/aout.html 2/

2016/10/10 Neural networks and deep learning 欧拉的博客:www.liuhao.me usually just elementary algebra and plots of functions, which epect most readers will be okay with. occasionally use more advanced mathematics, but have structured the material so you can follow even if some mathematical details elude you. The one chapter which uses heavier mathematics etensively is Chapter , which requires a little multivariable calculus and linear algebra. f those arent familiar, begin Chapter with a discussion of how to navigate the mathematics. f youre finding it really heavy going, you can simply skip to the summary of the chapters main results. n any case, theres no need to worry about this at the outset. ts rare for a book to aim to be both principle-oriented and hands- on. ut believe youll learn best if we build out the fundamental ideas of neural networks. ell develop living code, not just abstract theory, code which you can eplore and etend. This way youll understand the fundamentals, both in theory and practice, and be well set to add further to your knowledge. ncecorpesecesoosceeseneureorsneepernn eernonress supern sorscenseunerreeoonsruononoercnporecensesens ourereeocopsrenuonsoounooseoureneresencoercusepese conce http://neuralnetworksanddeeplearning.com/aout.html /

2016/10/10 Neural networks and deep learning n te eerises and proles 欧拉的博客:www.liuhao.me ts not uncommon for technical books to include an admonition from the author that readers must do the eercises and problems. always feel a little peculiar when read such warnings. ill something bad happen to me if dont do the eercises and problems Of course not. ll gain some time, but at the epense of depth of understanding. ometimes thats worth it. ometimes its not. o whats worth doing in this book y advice is that you really should attempt most of the eercises, and you should aim not to do most of the problems. ou should do most of the eercises because theyre basic checks that youve understood the material. f you cant solve an eercise relatively easily, youve probably missed something fundamental. Of course, if you do get stuck on an occasional eercise, just move on - chances are its just a small misunderstanding on your part, or maybe ve worded something poorly. ut if most eercises are a struggle, then you probably need to reread some earlier material. The problems are another matter. Theyre more difficult than the eercises, and youll likely struggle to solve some problems. Thats annoying, but, of course, patience in the face of such frustration is the only way to truly understand and internalie a subject. ith that said, dont recommend working through all the problems. hats even better is to find your own project. aybe you want to use neural nets to classify your music collection. Or to predict stock prices. Or whatever. ut ind a proet ou are aout. Then you can ignore the problems in the book, or use them simply as inspiration for work on your own project. truggling with a project you care about will teach you far more than working through any number of set problems. motional commitment is a key to achieving mastery. Of course, you may not have such a project in mind, at least up front. Thats fine. ork through those problems you feel motivated to work on. nd use the material in the book to help you search for ideas for creative personal projects. Neural Networks and Deep earning hat this book is about On the eercises and problems sing neural nets to recognie handwritten digits ow the backpropagation algorithm works mproving the way neural networks learn visual proof that neural nets can compute any function hy are deep neural networks hard to train Deep learning ppendi: s there a siple algorithm for intelligence cknowledgements Frequently sked uestions f you benefit from the book, please make a small donation. suggest , but you can choose the amount. Sponsors Thanks to all the supporters who made the book possible, with especial thanks to avel Dudrenov. Thanks also to all the contributors to the ugfinder all of Fame. Resources ook F Code repository ichael Nielsens project announcement mailing list Deep earning, draft book in preparation, by oshua engio, an oodfellow, and aron Courville http://neuralnetworksanddeeplearning.com/exercisesandprolems.html 1/2

资料库

Neural Networks and Deep Learning.pdf

相关推荐

课程资源

热门标签

最新资料