Buch, Englisch, 145 Seiten, Format (B × H): 155 mm x 235 mm
A 12-Week Introductory Course
Buch, Englisch, 145 Seiten, Format (B × H): 155 mm x 235 mm
Reihe: Studies in Computational Intelligence
ISBN: 978-3-032-44920-7
Verlag: Springer
This introductory text builds a bridge between mathematical foundations and machine learning algorithms. Assuming only a basic background in mathematics, statistics, and scientific computing, it develops the key concepts underlying machine learning through three fundamental questions: Which samples are important? Which features are important? How can we generate new samples? Topics include support vector machines, sparse learning, optimization, information theory, and generative models.
Emphasizing mathematical understanding rather than black-box use of software, the book is suitable for advanced undergraduate students, beginning graduate students, and practitioners seeking a rigorous introduction to machine learning.
Zielgruppe
Graduate
Autoren/Hrsg.
Fachgebiete
Weitere Infos & Material
1 Introduction.- 2 Similarity, distance, convexity,.- 2.1 Introduction.- 2.2 Distances and similarity.- 2.3 Why Euclidean distance?.- 2.4 Convex set.- 2.5 Convex function.- Problems.- References.- 3 Optimization, duality, regularization.- 3.1 Introduction.- 3.2 Convex optimization.- 3.3 A schema of iterative optimization.- 3.4 Duality.- 3.5 Regularization.- 3.6 *Algorithms and Convergence.- Problems.- References.- 4 Which samples are important?.- 4.1 Introduction.- 4.2 Linearly separable classes.- 4.3 Non-linear separable classes: penalization.- 4.4 Non-linear separable classes: kernel method.- 4.5 SVM regression.- 4.6 Answer to the first question.- Problems.- References.- 5 Which features are important?.- 5.1 Introduction.- 5.2 Regularized linear regression.- 5.3 Ridge regression.- 5.4 Sparse regression.- 5.5 Sparsity revealed.- 5.6 Answer to the second question.- Problems.- References.- 6 Can we generate it?.- 6.1 Introduction.- 6.2 Markov and Shannon.- 6.3 Shannon’s measure of information.- 6.4 Connection to thermodynamics.- 6.5 Learning probability distribution.- 6.5.1 Independence hypothesis.- 6.5.2 Mixture hypothesis.- 6.6 Energy-based models.- 6.6.1 EBM for classification or regression.- 6.6.2 EBM for generation.- 6.7 Answer to the third question.- Problems.- References.- Solutions.- Index.




