Probabilistic Graphical Models
Daphne Koller is a professor in Stanford University's Department of Computer Science. Nir Friedman is a professor of Hebrew University's Department of Computer Science and Engineering.
Most tasks necessitate the use of reasoning—the ability to reach conclusions based on available data. This book's framework of probabilistic graphical models provides a broad approach to this issue. The approach is model-based, allowing for the creation of interpretable models that can subsequently be changed by reasoning algorithms. These models can also be learnt automatically from data, allowing the approach to be employed in situations where manually building a model would be challenging, if not impossible. Because uncertainty is an unavoidable aspect of most real-world applications, the book focuses on probabilistic models, which make the uncertainty explicit and provide more accurate models.
Bayesian networks, undirected Markov networks, discrete and continuous models, and adaptations to deal with dynamical systems and relational data are all covered in Probabilistic Graphical Models. The work discusses the three main cornerstones of each model class: representation, inference, and learning, offering both basic principles and advanced techniques. Finally, the book considers the proposed framework's application to causal reasoning and decision making under uncertainty. Each chapter's primary material gives a deep technical development of the key themes. Most chapters also include additional material in the form of skill boxes, which describe techniques; case study boxes, which discuss empirical cases related to the approach described in the text, such as applications in computer vision, robotics, natural language understanding, and computational biology; and concept boxes, which present significant concepts drawn from the chapter's material. To meet their specific needs, instructors (and readers) can group chapters in various combinations ranging from core topics to more technically advanced material.
Author: Daphne Koller and Nir Friedman
Link to buy: https://www.amazon.com/Probabilistic-Graphical-Models-Principles-Computation/dp/0262013193/
Ratings: 4.4 out of 5 stars (from 104 reviews)
Best Sellers Rank: #124,440 in Books
#16 in Computer Vision & Pattern Recognition
#23 in Natural Language Processing (Books)
#25 in Machine Theory (Books)
