Bayes, AI, and Deep Learning
Foundations of Data Science
1st Edition
Will be published approx. on 17. November 2026
Book
Hardback
818 pages
978-1-032-47122-8 (ISBN)
Description
From Alan Turing breaking the Enigma code to ChatGPT reshaping how we work and create, artificial intelligence has always been powered by a single unifying principle: learning from data under uncertainty. This book reveals the mathematical machinery behind modern AI, weaving together Bayesian probability, statistical learning, and deep neural networks into a coherent intellectual framework. Whether you seek to understand why large language models hallucinate, how recommendation systems predict your preferences, or what makes reinforcement learning agents master complex games, this book equips you with both the theoretical foundations and practical intuitions that define the modern AI playbook.
Key Features:
Builds Bayesian reasoning from first principles through historical examples like submarine search and WWII code-breaking
Bridges classical statistics and deep learning, connecting linear regression to transformers
Covers the complete AI stack: probability, decision theory, Gaussian processes, neural networks, CNNs, NLP, and LLMs
Emphasizes uncertainty quantification-building systems that know what they don't know
Practical applications across finance, healthcare, operations, and autonomous systems
Emerging topics: AI agents, RLHF, and retrieval-augmented generation
Written for graduate students, data scientists, and quantitatively-minded practitioners, this book assumes comfort with calculus and basic probability while building sophisticated intuitions progressively. Business analysts will appreciate the decision-theoretic framing; engineers will value the architectural insights; researchers will find rigorous foundations for further study. Whether used as a course textbook, professional reference, or intellectual companion for understanding the AI revolution transforming every industry, this book delivers the rare combination of mathematical depth and accessible exposition that makes complex ideas genuinely understandable.
Key Features:
Builds Bayesian reasoning from first principles through historical examples like submarine search and WWII code-breaking
Bridges classical statistics and deep learning, connecting linear regression to transformers
Covers the complete AI stack: probability, decision theory, Gaussian processes, neural networks, CNNs, NLP, and LLMs
Emphasizes uncertainty quantification-building systems that know what they don't know
Practical applications across finance, healthcare, operations, and autonomous systems
Emerging topics: AI agents, RLHF, and retrieval-augmented generation
Written for graduate students, data scientists, and quantitatively-minded practitioners, this book assumes comfort with calculus and basic probability while building sophisticated intuitions progressively. Business analysts will appreciate the decision-theoretic framing; engineers will value the architectural insights; researchers will find rigorous foundations for further study. Whether used as a course textbook, professional reference, or intellectual companion for understanding the AI revolution transforming every industry, this book delivers the rare combination of mathematical depth and accessible exposition that makes complex ideas genuinely understandable.
More details
Series
Language
English
Place of publication
United Kingdom
Publishing group
Taylor & Francis Ltd
Target group
College/higher education
Professional and scholarly
Academic, Postgraduate, Undergraduate Advanced, and Undergraduate Core
Illustrations
185 farbige Zeichnungen, 131 s/w Tabellen, 185 farbige Abbildungen
131 Tables, black and white; 185 Line drawings, color; 185 Illustrations, color
Dimensions
Height: 254 mm
Width: 178 mm
Weight
453 gr
ISBN-13
978-1-032-47122-8 (9781032471228)
Copyright in bibliographic data and cover images is held by Nielsen Book Services Limited or by the publishers or by their respective licensors: all rights reserved.
Schweitzer Classification
Other editions
Additional editions
Book
approx. 11/2026
1st Edition
CRC Press
€103.07
Not yet published
Persons
Vadim Sokolov is Associate Professor in the Department of Systems Engineering and Operations Research at George Mason University, where he develops Bayesian methods, machine learning algorithms, and deep learning architectures for complex systems. His research spans statistical learning theory, probabilistic modeling, and intelligent transportation systems, with publications in leading journals including Bayesian Analysis, Transportation Research, and IEEE Transactions on Intelligent Transportation Systems.
Before joining Mason in 2016, Sokolov served as Visiting Assistant Professor of Statistics at the University of Chicago Booth School of Business and Principal Computational Scientist at Argonne National Laboratory, where he led the development of POLARIS, a large-scale agent-based transportation simulation framework, and the GREET life-cycle analysis model used by over 800 organizations worldwide.
Sokolov earned his Ph.D. in Computational Mathematics from Northern Illinois University and holds a diploma in Applied Mathematics with High Honors from Rostov State University, Russia. His work bridges rigorous statistical foundations with practical applications in energy systems, urban analytics, and data-driven decision-making. He is a member of INFORMS, the International Society for Bayesian Analysis, and the American Statistical Association.
Nicholas Polson is the Robert Law, Jr. Professor of Econometrics and Statistics at the University of Chicago Booth School of Business, where he has shaped modern Bayesian statistics and machine learning since 1991. A leading authority on probabilistic modeling, his research encompasses Markov chain Monte Carlo methods, particle learning, financial econometrics, and deep learning theory, with foundational contributions to stochastic volatility modeling, sparse Bayesian estimation, and high-dimensional inference.
Polson's influential work includes developing particle filtering algorithms for sequential learning and Bayesian regularization methods ranging from Tikhonov to horseshoe priors. His article "Bayesian Analysis of Stochastic Volatility Models" was recognized as one of the most influential papers in the 20th anniversary issue of the Journal of Business and Economic Statistics. He co-authored AIQ: How People and Machines Are Smarter Together (2018), exploring the synergy between human intelligence and artificial intelligence.
Polson earned his master's degree with First Class Honours from Worcester College, Oxford University, and his Ph.D. from the University of Nottingham. His work bridges theoretical foundations in probability and statistics with practical applications in finance, forecasting, and data science, establishing him as a pioneer in connecting classical Bayesian methods to modern deep learning.
Before joining Mason in 2016, Sokolov served as Visiting Assistant Professor of Statistics at the University of Chicago Booth School of Business and Principal Computational Scientist at Argonne National Laboratory, where he led the development of POLARIS, a large-scale agent-based transportation simulation framework, and the GREET life-cycle analysis model used by over 800 organizations worldwide.
Sokolov earned his Ph.D. in Computational Mathematics from Northern Illinois University and holds a diploma in Applied Mathematics with High Honors from Rostov State University, Russia. His work bridges rigorous statistical foundations with practical applications in energy systems, urban analytics, and data-driven decision-making. He is a member of INFORMS, the International Society for Bayesian Analysis, and the American Statistical Association.
Nicholas Polson is the Robert Law, Jr. Professor of Econometrics and Statistics at the University of Chicago Booth School of Business, where he has shaped modern Bayesian statistics and machine learning since 1991. A leading authority on probabilistic modeling, his research encompasses Markov chain Monte Carlo methods, particle learning, financial econometrics, and deep learning theory, with foundational contributions to stochastic volatility modeling, sparse Bayesian estimation, and high-dimensional inference.
Polson's influential work includes developing particle filtering algorithms for sequential learning and Bayesian regularization methods ranging from Tikhonov to horseshoe priors. His article "Bayesian Analysis of Stochastic Volatility Models" was recognized as one of the most influential papers in the 20th anniversary issue of the Journal of Business and Economic Statistics. He co-authored AIQ: How People and Machines Are Smarter Together (2018), exploring the synergy between human intelligence and artificial intelligence.
Polson earned his master's degree with First Class Honours from Worcester College, Oxford University, and his Ph.D. from the University of Nottingham. His work bridges theoretical foundations in probability and statistics with practical applications in finance, forecasting, and data science, establishing him as a pioneer in connecting classical Bayesian methods to modern deep learning.
Content
1 Probability and Uncertainty. 2 Bayes Rule. 3 Bayesian Learning. 4 Utility, Risk and Decisions. 5 A/B Testing. 6 Bayesian Hypothesis Testing. 7 Stochastic Processes. 8 Gaussian Processes. 9 Reinforcement Learning. 10 Unreasonable Effectiveness of Data. 11 Pattern Matching. 12 Linear Regression. 13 Logistic Regression and Generalized Linear Models. 14 Tree Models. 15 Forecasting. 16 Model Selection. 17 Statistical Learning Theory and Regularization. 18 Neural Networks. 19 Theory of Deep Learning. 20 Gradient Descent. 21 Quantile Neural Networks. 22 Convolutional Neural Networks. 23 Natural Language Processing. 24 Large Language Models. 25 AI Agents.