Muutke küpsiste eelistusi

E-raamat: Probabilistic Deep Learning: With Python, Keras and TensorFlow Probability

4.33/5 (36 hinnangut Goodreads-ist)
, ,
  • Formaat: 296 pages
  • Ilmumisaeg: 11-Oct-2020
  • Kirjastus: Manning Publications
  • Keel: eng
  • ISBN-13: 9781638350408
Teised raamatud teemal:
  • Formaat - EPUB+DRM
  • Hind: 43,42 €*
  • * hind on lõplik, st. muud allahindlused enam ei rakendu
  • Lisa ostukorvi
  • Lisa soovinimekirja
  • See e-raamat on mõeldud ainult isiklikuks kasutamiseks. E-raamatuid ei saa tagastada.
Probabilistic Deep Learning: With Python, Keras and TensorFlow ProbabilitySuurem pilt
  • Formaat: 296 pages
  • Ilmumisaeg: 11-Oct-2020
  • Kirjastus: Manning Publications
  • Keel: eng
  • ISBN-13: 9781638350408
Teised raamatud teemal:

DRM piirangud

  • Kopeerimine (copy/paste):

    ei ole lubatud

  • Printimine:

    ei ole lubatud

  • Kasutamine:

    Digitaalõiguste kaitse (DRM)
    Kirjastus on väljastanud selle e-raamatu krüpteeritud kujul, mis tähendab, et selle lugemiseks peate installeerima spetsiaalse tarkvara. Samuti peate looma endale  Adobe ID Rohkem infot siin. E-raamatut saab lugeda 1 kasutaja ning alla laadida kuni 6'de seadmesse (kõik autoriseeritud sama Adobe ID-ga).

    Vajalik tarkvara
    Mobiilsetes seadmetes (telefon või tahvelarvuti) lugemiseks peate installeerima selle tasuta rakenduse: PocketBook Reader (iOS / Android)

    PC või Mac seadmes lugemiseks peate installima Adobe Digital Editionsi (Seeon tasuta rakendus spetsiaalselt e-raamatute lugemiseks. Seda ei tohi segamini ajada Adober Reader'iga, mis tõenäoliselt on juba teie arvutisse installeeritud )

    Seda e-raamatut ei saa lugeda Amazon Kindle's. 

Probabilistic Deep Learning shows how probabilistic deep learning models gives readers the tools to identify and account for uncertainty and potential errors in their results.

 

Starting by applying the underlying maximum likelihood principle of curve fitting to deep learning, readers will move on to using the Python-based Tensorflow Probability framework, and set up Bayesian neural networks that can state their uncertainties.

 

Key Features

·   The maximum likelihood principle that underlies deep learning applications

·   Probabilistic DL models that can indicate the range of possible outcomes

·   Bayesian deep learning that allows for the uncertainty occurring in real-world situations

·   Applying probabilistic principles to variational auto-encoders

 

Aimed  at  a  reader  experienced  with  developing  machine  learning  or deep learning applications.

 

About the technology

Probabilistic deep learning models are better suited to dealing with the noise  and  uncertainty  of  real  world  data a  crucial  factor  for self-driving cars, scientific results, financial industries, and other accuracy-critical applications.

 

Oliver Dürr is professor for data science at the University of Applied Sciences in Konstanz, Germany.

 

Beate Sick holds a chair for applied statistics at ZHAW, and works as a researcher and lecturer at the University of Zurich, and as a lecturer at ETH Zurich.

 

Elvis Murina is a research assistant, responsible for the extensive exercises that accompany this book.

 

Dürr and Sick are both experts in machine learning and statistics. They have supervised numerous bachelors, masters, and PhD the seson the topic of deep learning, and planned and conducted several postgraduate and masters-level deep learning courses. All three authors have been working with deep learning methods since 2013 and have extensive experience in both teaching the topic and developing probabilistic deep learning models.
Preface xi
Acknowledgments xii
About this book xiv
About the authors xvii
About the cover illustration xviii
PART 1 BASICS OF DEEP LEARNING
1(90)
1 Introduction to probabilistic deep learning
3(22)
1.1 A first look at probabilistic models
4(2)
1.2 A first brief look at deep learning (DL)
6(2)
A success story
8(1)
1.3 Classification
8(8)
Traditional approach to image classification
9(3)
Deep learning approach to image classification
12(2)
Non-probabilistic classification
14(1)
Probabilistic classification
14(2)
Bayesian probabilistic classification
16(1)
1.4 Curve fitting
16(5)
Non-probabilistic curve fitting
17(1)
Probabilistic curve fitting
18(2)
Bayesian probabilistic curve fitting
20(1)
1.5 When to use and when not to use DL?
21(2)
When not to use DL
21(1)
When to use DL
22(1)
When to use and when not to use probabilistic models
22(1)
1.6 What you'll learn in this book
23(2)
2 Neural network architectures
25(37)
2.1 Fully connected neural networks (fcNNs)
26(18)
The biology that inspired the design of artificial NNs
26(2)
Getting started with implementing an NN
28(10)
Using a fully connected NN (fcNN) to classify images
38(6)
2.2 Convolutional NNs for image-like data
44(12)
Main ideas in a CNN architecture
44(3)
A minimal CNN for edge lovers
47(3)
Biological inspiration for a CNN architecture
50(2)
Building and understanding a CNN
52(4)
2.3 One-dimensional CNNs for ordered data
56(6)
Format of time-ordered data
57(1)
What's special about ordered data?
58(1)
Architectures for time-ordered data
59(3)
3 Principles of curve fitting
62(29)
3.1 "Hello world" in curve fitting
63(6)
Fitting a linear regression model based on a loss function
65(4)
3.2 Gradient descent method
69(9)
Loss with one free model parameter
69(4)
Loss with two free model parameters
73(5)
3.3 Special DL sauce
78(2)
Mini-batch gradient descent
78(1)
Using SGD variants to speed up the learning
79(1)
Automatic differentiation
79(1)
3.4 Backpropagation in DL frameworks
80(11)
Static graph frameworks
81(7)
Dynamic graph frameworks
88(3)
PART 2 MAXIMUM LIKELIHOOD APPROACHES FOR PROBABILISTIC DL MODELS
91(106)
4 Building loss functions with the likelihood approach
93(35)
4.1 Introduction to the MaxLike principle: The mother of all loss functions
94(5)
4.2 Deriving a loss function for a classification problem
99(12)
Binary classification problem
99(6)
Classification problems with more than two classes
105(4)
Relationship between NLL, cross entropy, and Kullback-Leibler divergence
109(2)
4.3 Deriving a loss function for regression problems
111(17)
Using a NN without hidden layers and one output neuron for modeling a linear relationship between input and output
111(8)
Using a NN with hidden layers to model non-linear relationships between input and output
119(2)
Using an NN with additional output for regression tasks with nonconstant variance
121(7)
5 Probabilistic deep learning models with Tensor Flow Probability
128(29)
5.1 Evaluating and comparing different probabilistic prediction models
130(2)
5.2 Introducing TensorFlow Probability (TFP)
132(3)
5.3 Modeling continuous data with TFP
135(10)
Fitting and evaluating a linear regression model with constant variance
136(4)
Fitting and evaluating a linear regression model with a nonconstant standard deviation
140(5)
5.4 Modeling count data with TensorFlow Probability
145(12)
The Poisson distribution for count data
148(5)
Extending the Poisson distribution to a zero-inflated Poisson (TIP) distribution
153(4)
6 Probabilistic deep learning models in the wild
157(40)
6.1 Flexible probability distributions in state-of-the-art DL models
159(6)
Multinomial distribution as a flexible distribution
160(2)
Making sense of discretized logistic mixture
162(3)
6.2 Case study: Bavarian roadkills
165(1)
6.3 Go with the flow: Introduction to normalizing flows (NFs)
166(31)
The principle idea of NFs
168(2)
The change of variable technique for probabilities
170(5)
Fitting an NF to data
175(2)
Going deeper by chaining flows
177(4)
Transformation between higher dimensional spaces
181(2)
Using networks to control flows
183(5)
Fun with flows: Sampling faces
188(9)
PART 3 BAYESIAN APPROACHES FOR PROBABILISTIC
7 Bayesian learning
197(32)
7.1 What's wrong with non-Bayesian DL: The elephant in the room
198(3)
7.2 The first encounter with a Bayesian approach
201(6)
Bayesian model: The hacker's way
202(4)
What did we just do?
206(1)
7.3 The Bayesian approach for probabilistic models
207(22)
Training and prediction with a Bayesian model
208(16)
A coin toss as a Hello World example for Bayesian models 213* Revisiting the Bayesian linear regression model
224(5)
8 Bayesian neural networks
229(35)
8.1 Bayesian neural networks (BNNs)
230(2)
8.2 Variational inference (VI) as an approximative Bayes approach
232(11)
Looking under the hood of VI*
233(5)
Applying VI to the toy problem*
238(5)
8.3 Variational inference with TensorFlow Probability
243(2)
8.4 MC dropout as an approximate Bayes approach
245(7)
Classical dropout used during training
246(3)
MC dropout used during train and test times
249(3)
8.5 Case studies
252(12)
Regression case study on extrapolation
252(4)
Classification case study with novel classes
256(8)
Glossary of terms and abbreviations 264(5)
Index 269
Oliver Dürr is professor for data science at the University of Applied Sciences in Konstanz, Germany.

 

Beate Sick holds a chair for applied statistics at ZHAW, and works as a researcher and lecturer at the University of Zurich, and as a lecturer at ETH Zurich.

 

Elvis Murina is a research assistant, responsible for the extensive exercises that accompany this book.

 

Dürr and Sick are both experts in machine learning and statistics. They have supervised numerous bachelors, masters, and PhD the seson the topic of deep learning, and planned and conducted several postgraduate and masters-level deep learning courses. All three authors have been working with deep learning methods since 2013 and have extensive experience in both teaching the topic and developing probabilistic deep learning models.
Lisainfo e-raamatute kohta Lisainfo e-raamatute kohta
Püsilink: https://www.kriso.ee/db/97816383504086e.html
Märksõnad: