The Handbook of Brain Theory and Neural Networks

Rating:
Summary: Misleading title, a useful book otherwise
Review: Look through this book to convince yourself that an exact brain theory does not exist. The arrangement of the articles by the first letter of their title tells it all (consider classifying animals by the first letter of their name). The editors wrongly assume that mathematical methods equal theory; actually, theory is a small conceptual tent under which a large number of experimentally established facts can be gathered. In most cases, mathematics is a very useful tool in pitching this tent, but it has little to do with the tent itself.

An exact theory of the brain may be possible and actually we are in dire need of it. Unfortunately, nobody has come up with it yet. This book is an encyclopedia of various mathematical methods that have been used to solve various neuroscience problems. These methods and solutions are as diverse as the problems themselves. Don't look for common themes in this book. If you are looking for a unified brain theory, you'd be much better off reading standard neuroscience textbooks. I do hope one day we'll be able to cast these vague ideas into something precise and, most likely, mathematical. Sadly, not today. I own a copy of this book and use it to remind me why and how we have failed so far.

It should be kept in mind that it is not at all clear that "neural" networks can emulate consciousness. They may or they may not. Firstly, a single neuron resembles a computer processor in its complexity and is a constantly evolving entity. Secondly, only 10% of brain cells are neurons and the remaining 90% (glial cells) now too appear to be involved in information processing. At a more fundamental level, consciousness may be less algorithmic and computational than we expect. Finally, the brain and the reality "outside the brain" are a two-way street. As the great neuroscientist Cajal put it, "As long as our brain remains an arcanum, the Universe, a reflection of its structure, will also be a mystery". If we assume the brain analyzes something, we need to define a reality independent of this analysis -- a hardly possible task if standard "input-output" approaches are used.

If the title of this book were "Current Mathematical Methods in Neurosciences", I'd have no problem giving it five stars.

Rating:
Summary: Basic science for consciousness
Review: Research is tedious, but if you want to know the nitty-gritty of mind-brain theory and neural networking, this book is an invaluable resource for basic, relevant, and accessible papers on the subjects. Encompassing seminal works from an unusually broad range of disciplines, here is an outstanding reference for those concerned with the mechanisms of intelligence.

Rating:
Summary: Excellent compilation
Review: This complilation of articles by leading experts in the field gives an excellent overview of studies in cognitive theory and the theory and applications of neural networks. The first two parts of the book give an overview and background of the properties of neurons and gives guidance to the reader on what sequence the articles are to be read. I did not read all of the articles, but only those that piqued my interest. I found the following articles particularly well-written and informative: 1. "Applications of Neural Networks": Outlines the diverse applications of neural networks to signal processing, time series, imaging, etc. 2. "Astronomy": Neural network applications in astronomy, such as adaptive optics and telescope guidance. 3. "Chains of Coupled Oscillators": Their connection with the lamprey central pattern generator. 4. "Chaos in Axons": An excellent review of chaos experimentally in squid axons and numerically with nerve equations. 5. "Collective Behavior of Coupled Oscillators": A study of the phase and complex Ginzburg-Landau model. 6. "Computer Modeling Methods for Neurons": Good overview of numerical modeling of neurons. 7. "Computing with Attractors": Overview of omputing and feedback networks with attractors and a fascinating discussion of the possible existence of attractors in the brain. 8. "Constrained Optimization and the Elastic Net": Useful discussion of application of neural networks to optimization problems. 9. "Data Clustering and Learning": Good discussion of parameter estimation of mixture models by parametric statistics and vector quantization of a data set by combinatorial optimization. 10. "Diffusion Models of Neuron Activity": Discusses 1-dimensional stochastic diffusion models for the neuron membrane potential. 11. "Disease: Neural Network Models": Interesting overview of neural net computational models of various mental illnesses. 12. "Dynamics and Bifurcation of Neural Networks": Discussion of neural nets and their behavior as dynamical systems. 13. "Emotion and Computational Neuroscience": Fascinating discussion of computational models of emotion. 14. "Investment Management": A discussion of tactical asset allocation neural network methods in asset management. 15. "Learning and Centralization: Theoretical Bounds": Overview of computational learning theory. 16. "Locust Flight": Interesting neural network study of the locust flight system. 17. "Neural Optimization": Discussion of combinatorial optimization using Ising and Potts neural networks. 18. "PAC Learning and Neural Networks": Overview of the Valiant "probabilistically correct learning paradigm in neural networks. 19. "Protein Structure Prediction": Neural network applications to prediction of protein secondary structure. 20. "Schema Theory": Extremely interesting overview of schemas. 21. "Speech Recognition: Pattern Matching": Excellent discussion of the applications of hidden Markov models to speech recognition. 22. "Statistical Mechanics of Neural Networks": Discussion of the use of the Hopfield model in neural networks. 23. Vapnik-Chervonenkis Dimension of Neural Networks": Very interesting discussion of the VC-dimension of neural networks.

Rating:
Summary: Neural Network Bible
Review: This is THE neural network and brain theory reference. Owning it is like owning an entire library, though much more compact.

If you take a look at the table of contents, you'll see the massive value in this book. If you're into neural nets and brain theory, or want to be, you need this book.