Generative Deep Learning (Record no. 359829)
[ view plain ]
| 000 -LEADER | |
|---|---|
| fixed length control field | 09347 a2200181 4500 |
| 005 - DATE AND TIME OF LATEST TRANSACTION | |
| control field | 20241018161810.0 |
| 008 - FIXED-LENGTH DATA ELEMENTS--GENERAL INFORMATION | |
| fixed length control field | 241018b |||||||| |||| 00| 0 eng d |
| 020 ## - INTERNATIONAL STANDARD BOOK NUMBER | |
| International Standard Book Number | 9789355429988 |
| 041 ## - LANGUAGE CODE | |
| Language code of text/sound track or separate title | English |
| 100 ## - MAIN ENTRY--PERSONAL NAME | |
| Author | Foster D. |
| 245 ## - TITLE STATEMENT | |
| Title | Generative Deep Learning |
| Remainder of title | :Teaching Machines To Paint, Write, Compose And Play |
| 250 ## - EDITION STATEMENT | |
| Edition statement | 2nd |
| 260 ## - PUBLICATION, DISTRIBUTION, ETC. | |
| Name of publisher, distributor, etc. | SPD |
| Date of publication, distribution, etc. | 2023 |
| 300 ## - PHYSICAL DESCRIPTION | |
| Extent | 426 |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | Table of Contents<br/>Foreword. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv<br/>Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii<br/>Part I. Introduction to Generative Deep Learning<br/>1. Generative Modeling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3<br/>What Is Generative Modeling? 4<br/>Generative Versus Discriminative Modeling 5<br/>The Rise of Generative Modeling 6<br/>Generative Modeling and AI 8<br/>Our First Generative Model 9<br/>Hello World! 9<br/>The Generative Modeling Framework 10<br/>Representation Learning 12<br/>Core Probability Theory 15<br/>Generative Model Taxonomy 18<br/>The Generative Deep Learning Codebase 20<br/>Cloning the Repository 20<br/>Using Docker 21<br/>Running on a GPU 21<br/>Summary 21<br/>2. Deep Learning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23<br/>Data for Deep Learning 24<br/>Deep Neural Networks 25<br/>vii<br/>What Is a Neural Network? 25<br/>Learning High-Level Features 26<br/>TensorFlow and Keras 27<br/>Multilayer Perceptron (MLP) 28<br/>Preparing the Data 28<br/>Building the Model 30<br/>Compiling the Model 35<br/>Training the Model 37<br/>Evaluating the Model 38<br/>Convolutional Neural Network (CNN) 40<br/>Convolutional Layers 41<br/>Batch Normalization 46<br/>Dropout 49<br/>Building the CNN 51<br/>Training and Evaluating the CNN 53<br/>Summary 54<br/>Part II. Methods<br/>3. Variational Autoencoders. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59<br/>Introduction 60<br/>Autoencoders 61<br/>The Fashion-MNIST Dataset 62<br/>The Autoencoder Architecture 63<br/>The Encoder 64<br/>The Decoder 65<br/>Joining the Encoder to the Decoder 67<br/>Reconstructing Images 69<br/>Visualizing the Latent Space 70<br/>Generating New Images 71<br/>Variational Autoencoders 74<br/>The Encoder 75<br/>The Loss Function 80<br/>Training the Variational Autoencoder 82<br/>Analysis of the Variational Autoencoder 84<br/>Exploring the Latent Space 85<br/>The CelebA Dataset 85<br/>Training the Variational Autoencoder 87<br/>Analysis of the Variational Autoencoder 89<br/>Generating New Faces 90<br/>viii | Table of Contents<br/>Latent Space Arithmetic 91<br/>Morphing Between Faces 92<br/>Summary 93<br/>4. Generative Adversarial Networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95<br/>Introduction 96<br/>Deep Convolutional GAN (DCGAN) 97<br/>The Bricks Dataset 98<br/>The Discriminator 99<br/>The Generator 101<br/>Training the DCGAN 104<br/>Analysis of the DCGAN 109<br/>GAN Training: Tips and Tricks 110<br/>Wasserstein GAN with Gradient Penalty (WGAN-GP) 113<br/>Wasserstein Loss 114<br/>The Lipschitz Constraint 115<br/>Enforcing the Lipschitz Constraint 116<br/>The Gradient Penalty Loss 117<br/>Training the WGAN-GP 119<br/>Analysis of the WGAN-GP 121<br/>Conditional GAN (CGAN) 122<br/>CGAN Architecture 123<br/>Training the CGAN 124<br/>Analysis of the CGAN 126<br/>Summary 127<br/>5. Autoregressive Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129<br/>Introduction 130<br/>Long Short-Term Memory Network (LSTM) 131<br/>The Recipes Dataset 132<br/>Working with Text Data 133<br/>Tokenization 134<br/>Creating the Training Set 137<br/>The LSTM Architecture 138<br/>The Embedding Layer 138<br/>The LSTM Layer 140<br/>The LSTM Cell 142<br/>Training the LSTM 144<br/>Analysis of the LSTM 146<br/>Recurrent Neural Network (RNN) Extensions 149<br/>Stacked Recurrent Networks 149<br/>Table of Contents | ix<br/>Gated Recurrent Units 151<br/>Bidirectional Cells 153<br/>PixelCNN 153<br/>Masked Convolutional Layers 154<br/>Residual Blocks 156<br/>Training the PixelCNN 158<br/>Analysis of the PixelCNN 159<br/>Mixture Distributions 162<br/>Summary 164<br/>6. Normalizing Flow Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167<br/>Introduction 168<br/>Normalizing Flows 169<br/>Change of Variables 170<br/>The Jacobian Determinant 172<br/>The Change of Variables Equation 173<br/>RealNVP 174<br/>The Two Moons Dataset 174<br/>Coupling Layers 175<br/>Training the RealNVP Model 181<br/>Analysis of the RealNVP Model 184<br/>Other Normalizing Flow Models 186<br/>GLOW 186<br/>FFJORD 187<br/>Summary 188<br/>7. Energy-Based Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189<br/>Introduction 189<br/>Energy-Based Models 191<br/>The MNIST Dataset 192<br/>The Energy Function 193<br/>Sampling Using Langevin Dynamics 194<br/>Training with Contrastive Divergence 197<br/>Analysis of the Energy-Based Model 201<br/>Other Energy-Based Models 202<br/>Summary 203<br/>8. Diusion Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205<br/>Introduction 206<br/>Denoising Diffusion Models (DDM) 208<br/>The Flowers Dataset 208<br/>x | Table of Contents<br/>The Forward Diffusion Process 209<br/>The Reparameterization Trick 210<br/>Diffusion Schedules 211<br/>The Reverse Diffusion Process 214<br/>The U-Net Denoising Model 217<br/>Training the Diffusion Model 224<br/>Sampling from the Denoising Diffusion Model 225<br/>Analysis of the Diffusion Model 228<br/>Summary 231<br/>Part III. Applications<br/>9. Transformers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235<br/>Introduction 236<br/>GPT 236<br/>The Wine Reviews Dataset 237<br/>Attention 238<br/>Queries, Keys, and Values 239<br/>Multihead Attention 241<br/>Causal Masking 242<br/>The Transformer Block 245<br/>Positional Encoding 248<br/>Training GPT 250<br/>Analysis of GPT 252<br/>Other Transformers 255<br/>T5 256<br/>GPT-3 and GPT-4 259<br/>ChatGPT 260<br/>Summary 264<br/>10. Advanced GANs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267<br/>Introduction 268<br/>ProGAN 269<br/>Progressive Training 269<br/>Outputs 276<br/>StyleGAN 277<br/>The Mapping Network 278<br/>The Synthesis Network 279<br/>Outputs from StyleGAN 280<br/>StyleGAN2 281<br/>Table of Contents | xi<br/>Weight Modulation and Demodulation 282<br/>Path Length Regularization 283<br/>No Progressive Growing 284<br/>Outputs from StyleGAN2 286<br/>Other Important GANs 286<br/>Self-Attention GAN (SAGAN) 286<br/>BigGAN 288<br/>VQ-GAN 289<br/>ViT VQ-GAN 292<br/>Summary 294<br/>11. Music Generation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297<br/>Introduction 298<br/>Transformers for Music Generation 299<br/>The Bach Cello Suite Dataset 300<br/>Parsing MIDI Files 300<br/>Tokenization 303<br/>Creating the Training Set 304<br/>Sine Position Encoding 305<br/>Multiple Inputs and Outputs 307<br/>Analysis of the Music-Generating Transformer 309<br/>Tokenization of Polyphonic Music 313<br/>MuseGAN 317<br/>The Bach Chorale Dataset 317<br/>The MuseGAN Generator 320<br/>The MuseGAN Critic 326<br/>Analysis of the MuseGAN 327<br/>Summary 329<br/>12. World Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331<br/>Introduction 331<br/>Reinforcement Learning 332<br/>The CarRacing Environment 334<br/>World Model Overview 336<br/>Architecture 336<br/>Training 338<br/>Collecting Random Rollout Data 339<br/>Training the VAE 341<br/>The VAE Architecture 341<br/>Exploring the VAE 343<br/>Collecting Data to Train the MDN-RNN 346<br/>xii | Table of Contents<br/>Training the MDN-RNN 346<br/>The MDN-RNN Architecture 347<br/>Sampling from the MDN-RNN 348<br/>Training the Controller 348<br/>The Controller Architecture 349<br/>CMA-ES 349<br/>Parallelizing CMA-ES 351<br/>In-Dream Training 353<br/>Summary 356<br/>13. Multimodal Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359<br/>Introduction 360<br/>DALL.E 2 361<br/>Architecture 362<br/>The Text Encoder 362<br/>CLIP 362<br/>The Prior 367<br/>The Decoder 369<br/>Examples from DALL.E 2 373<br/>Imagen 377<br/>Architecture 377<br/>DrawBench 378<br/>Examples from Imagen 379<br/>Stable Diffusion 380<br/>Architecture 380<br/>Examples from Stable Diffusion 381<br/>Flamingo 381<br/>Architecture 382<br/>The Vision Encoder 382<br/>The Perceiver Resampler 383<br/>The Language Model 385<br/>Examples from Flamingo 388<br/>Summary 389<br/>14. Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391<br/>Timeline of Generative AI 392<br/>2014–2017: The VAE and GAN Era 394<br/>2018–2019: The Transformer Era 394<br/>2020–2022: The Big Model Era 395<br/>The Current State of Generative AI 396<br/>Large Language Models 396<br/>Table of Contents | xiii<br/>Text-to-Code Models 400<br/>Text-to-Image Models 402<br/>Other Applications 405<br/>The Future of Generative AI 407<br/>Generative AI in Everyday Life 407<br/>Generative AI in the Workplace 409<br/>Generative AI in Education 410<br/>Generative AI Ethics and Challenges 411<br/>Final Thoughts 413<br/>Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417<br/>xiv | Table of Contents |
| 700 ## - ADDED ENTRY--PERSONAL NAME | |
| Personal name | Friston K. |
| 942 ## - ADDED ENTRY ELEMENTS (KOHA) | |
| Koha item type | Books |
| Withdrawn status | Lost status | Source of classification or shelving scheme | Damaged status | Not for loan | Home library | Current library | Date acquired | Vendor | Net Price | Total Checkouts | Full call number | Barcode | Date last seen | Actual Price | Bill Date | Koha item type |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Dewey Decimal Classification | Cummins College of Engineering for Women Pune | Cummins College of Engineering for Women Pune | 14/10/2024 | 115 | 1350.00 | 006.32 FOS | CCEP-BK-67505 | 14/10/2024 | 1800.00 | 14/10/2024 | Books |
