Machine Learning: A Constraint-Based Approach
Morgan Kaufmann, 20 nov 2017 - 580 pagine
Machine Learning: A Constraint-Based 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 warm-up exercises that lead to deeper research problems. A software simulator is also included.
Cosa dicono le persone - Scrivi una recensione
Some years ago Marco Gori described to me his idea of constraint-based learning. Not much time was required to convince myself that Marco’s point of view was strongly innovative. I enthusiastically accepted his invitation to work together on some aspects related to this subject and so I could give a small contribution to Marco’s “big picture”. When I first saw the monograph, I was impressed by how much work Marco did in recent years on this topic and how clearly he can now present constraint-based learning. The book looks to me like a great “fresco”, in which a large variety of colours and tones have been put together in such a way as to create an exciting overall view. Despite the novelty of the approach described and the fact that research on this subject is quite recent, the material is presented with the elegance and clarity of a classical, well-established theory. Marco has the capability to guide through the various chapters of his "novel", so that the reader can enjoy the book like an enthusiastic journey in the world of constraints. Each concept and each theorem are presented in such a way that they seem to arise in the most natural way and exactly at the right moment. Although some advanced mathematical tools are used, rigour is achieved avoiding heavy formalism and allowing also less technically-oriented readers to enjoy the book. Summing up, this is an excellent monograph dealing with machine learning from an innovative point of view, which lays the foundations of a new theory with strong impact on machine learning applications. It opens the doors to a wide range of research directions. Definitely, Marco’s work is a “must” for researchers and practitioners in machine learning.
An excellent book on machine learning, with a rigorous yet not heavy formalism, capable of elaborating a unifying view of this discipline from a novel and interesting perspective. A stimulating reading both for students that want to approach this research field, and for experts that want to gain a better understanding of selected topics.
3 Linear Threshold Machines
4 Kernel Machines
5 Deep Architectures
6 Learning and Reasoning With Constraints
8 Answers to Exercises
C1 Functionals and Variations