Machine Learning: A ConstraintBased ApproachMachine Learning: A ConstraintBased Approach provides readers with a refreshing look at the basic models and algorithms of machine learning, with an emphasis on current topics of interest that includes neural networks and kernel machines. The book presents the information in a truly unified manner that is based on the notion of learning from environmental constraints. While regarding symbolic knowledge bases as a collection of constraints, the book draws a path towards a deep integration with machine learning that relies on the idea of adopting multivalued logic formalisms, like in fuzzy systems. A special attention is reserved to deep learning, which nicely fits the constrained based approach followed in this book. This book presents a simpler unified notion of regularization, which is strictly connected with the parsimony principle, and includes many solved exercises that are classified according to the Donald Knuth ranking of difficulty, which essentially consists of a mix of warmup exercises that lead to deeper research problems. A software simulator is also included.

What people are saying  Write a review
The book develops a new point of view of Machine Learning that is able to encompass different techniques and settings.
By seeing the problem of learning as that of satisfying constraints, the author builds a framework that
integrates symbolic and subsymbolic learning, one of the longstanding open problems in Artificial Intelligence.
At the same time, supervised and unsupervised learning, kernel machines and deep learning
are also modeled by the framework and new opportunities
arise from their combination.
All of this is discussed in the context of the problem of an agent acting in the evnrinoment, receiving feedback
and operating over time, showing how a constraint point of view can bring new insights on the problem.
The book is concerned with providing a big picture of Machine Learning, able to encompass current research directions, rather than
focusing on describing the details of each and every technique. As such, it is a mustread for researchers working in the field but
its accessiility makes it a good text for students and for people that want to get a fresh point of view on Machine Learning.
Contents
60  
3 Linear Threshold Machines  122 
4 Kernel Machines  186 
5 Deep Architectures  236 
6 Learning and Reasoning With Constraints  340 
7 Epilogue  446 
8 Answers to Exercises  452 
C1 Functionals and Variations  518 
C2 Basic Notion on Variations  520 
C3 EulerLagrange Equations  523 
C4 Variational Problems With Subsidiary Conditions  526 
534  
552  
Back Cover  561 