Buch, Englisch, 808 Seiten, Format (B × H): 210 mm x 280 mm
Buch, Englisch, 808 Seiten, Format (B × H): 210 mm x 280 mm
ISBN: 978-1-041-06625-5
Verlag: Taylor & Francis Ltd
This book provides a thorough introduction to Bayesian deep learning (BDL) BDL brings Bayesian reasoning to deep neural networks with the aim of making predictions useful for rational decision-making under uncertainty. Deep neural networks provide powerful and flexible models for this purpose, while Bayesian inference offers a principled framework for representing evidence and uncertainty and for assessing the decisions that predictions inform.
The work develops a broad view of BDL, combining a principal route through Bayesian methods with foundational and mathematical aspects, questions of scalability to contemporary models and datasets, applications, and topical developments. It places BDL within a wider perspective on machine learning: not as a specialised corner of the field, but as a framework for building powerful models while remaining formal about evidence, uncertainty, and the decisions that give predictions their purpose. Written by a large team of co-authors, the book covers a wide spectrum of methods in Bayesian deep learning, covering recent advances in the field.
A self-contained introduction to BDL, the book is perfect for students, researchers and practitioners interested in learning about, adopting or advancing Bayesian deep learning.
Zielgruppe
Academic, Postgraduate, Professional Practice & Development, and Undergraduate Advanced
Autoren/Hrsg.
Fachgebiete
- Mathematik | Informatik EDV | Informatik Daten / Datenbanken
- Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz Maschinelles Lernen
- Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz Computer Vision
- Mathematik | Informatik Mathematik Stochastik Bayesianische Inferenz
- Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz Neuronale Netzwerke
- Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz Mustererkennung, Biometrik
Weitere Infos & Material
Part I: Sampling methods
Chapter 1: Introduction to sampling
Chapter 2: Stochastic Gradient MCMC
Chapter 3: Simulation-based inference
Chapter 4: Sequential Monte Carlo samplers
Chapter 5: Low-precision sampling
Chapter 6: Discrete sampling
Part II: Laplace approximations
Chapter 7: Introduction
Chapter 8: Linearised Laplace
Chapter 9: Differential geometry
Chapter 10: Analytic functionality
Chapter 11: Model selection with the marginal likelihood
Part III: Variational inference
Chapter 12: Introduction to variational inference
Chapter 13: Gaussian mean-field variational inference
Chapter 14: Dropout-based variational inference
Chapter 15: Natural-gradient-based variational inference
Chapter 16: Particle-Based Variational Inference
Chapter 17: Variational inference in function space
Chapter 18: Sequential variational inference
Part IV: Ensemble methods
Chapter 19: Introduction to ensemble methods
Chapter 20: Hybrid ensemble sampling methods
Chapter 21: Practical applications of deep ensembles
Chapter 22: PAC-Bayesian deep neural network ensembles
Part V: Kernel methods
Chapter 23: Introduction to kernel methods
Chapter 24: Deep kernel learning
Chapter 25: Deep Gaussian processes
Chapter 26: Deep kernel processes and deep kernel machines
Chapter 27: Meta-learning deep kernel Gaussian processes
Part VI: Priors
Chapter 28: Introduction to priors
Chapter 29: Statistical properties of priors
Chapter 30: Informative priors
Chapter 31: Function-space priors
Part VII: Identifiability and symmetries
Chapter 32: Symmetries in Bayesian neural networks
Chapter 33: Symmetries and sampling
Chapter 34: Learning inductive bias and symmetries
Chapter 35: Interpretable inferences via Bayesian pyramids
Chapter 36: Bayesian mixture of experts in BDL
Part VIII: Scalability
Chapter 37: Scalable BDL by single forward pass inference
Chapter 38: Scalable Gaussian processes
Chapter 39: Last-layer inference
Chapter 40: Scalable Laplace approximations
Chapter 41: Learned Compression with Deterministic Codes
Chapter 42: Learned Compression with Stochastic Codes
Part IX: Applications
Chapter 43: Bayesian LLMs
Chapter 44: Confidence-based selective classification
Chapter 45: Biomedical data imputation
Chapter 46: Deep Bayesian spatio-temporal modelling of wildfires
Chapter 47: Software for BDL
Part X: Topical developments
Chapter 48: Subspace inference
Chapter 49: Implicitly Bayesian prediction rules
Chapter 50: Diffusion models
Chapter 51: Prior-data Fitted Networks
Chapter 52: Advances in singular learning theory for BDL
Chapter 53: Active Learning and Experimental Design
Chapter 54: Causal-aware BDL
Chapter 55: Credal BDL and Imprecise Probability in BDL
Chapter 56: Bayesian Reinforcement Learning
