Neural Networks Books

...This book gives an excellent view of the concept of ANN but covers them in pure mathemetical way. You can learn so much about the equations used to describe ANNs but I would doubt that you can learn enough to make you able to utilize them. It is a good read, but needs more practical approach to it, but talking about such a topic is not a well-defined path, so I give it a 4-star rating

This book opens up new possibilities. It includes a basically simple Graphical User Interface (GUI) that can be used in Microsoft Windows (and in fact uses it for the examples). I rate it with 4 stars instead of 5, because the instructions in the examples are for those experienced with NEURON. For beginners like myself, it would help to say which buttons should be clicked and which keys pressed.

This book describes the NEURON simulation system, which can be accessed for installation and instructions at the NEURON web site. Simulation implies using the realistic Hodgkin-Huxley neuron. NEURON was initially for individual neurons, but it has now been extended to networks.

For those who believe in the classical physical science of the 19th century, including physics, chemistry, thermodynamics, and the differential equations in which they are expressed, NEURON has a special meaning. The Hodgkin-Huxley neuron extended classical physical science to a wide range of neuron types and species. The reductionist work of Eric Kandel explained many types of synapses at the molecular level, and therefore explains the connection of neurons in a network in terms of classical physical science.

Our special interest is in networks of interneurons. The most accessible mammalian networks are those in the olfactory bulb of the rat. For this special class, classical physical science, using NEURON, extends into neurobiology. It DEFINES a physically possible network structure. It is likely that evolution will have exploited at least part of this structure to extend order. This possibility is there. It is real. And it is begging for study.

This work will not require a supercomputer. From the deterministic point of view of classical physical science, there is no magic in statistically large numbers of cells. Two dozen or less should be enough to display emerging order.

Freidemann Pulvermueller's "Neuroscience of Language" is an indispensible reference for anyone seeking more than cocktail-party knowledge about psychology, neurology, linguistics, and computational systems. Perhaps because it attempts to unite so many previously-unconnected fields in a concrete neural model of natural language (instead of in yet another abstract, mathematical model), the many technicalities of these fields unfold in a transparent, easy-to-follow order. The subtitle of Pulvermueller's book is not "Neuroscience for Dummies", but it succeeds in providing an integrated,lucid, instructive tour of the past 50 years of research in cognitive science.

Unfortunately, the book's subtitle is "On brain circuits of words and serial order", and when Pulvermueller finally turns to serial order in Chapter 8, I think his brain circuits start to malfunction. I say this because of an influential paper Karl Lashley wrote back in 1951, entitled, "On the problem of serial order in behavior". Briefly, the problem is that serial order isn't always serial (Lashley cited Spooner's toast "To our queer, old dean"). Unlike generations of pop pyschologists, Lashley's student, Noam Chomsky, understood this to be a devastating critique of behaviorism. The point about Spoonerisms is *not* that they are freaks of nature, but that they are ubiquitous in human behavior. Chomsky's prime example was the ubiquitous "transformation" of e.g. "John kissed Mary" into the passive voice "Mary was kissed by John", but almost every serial behavior in your human repertoire, from your route to work in the morning to the arpeggios you play on the piano, exhibits the serial non-seriality of Spoonerisms (or, more formally "metathesis"). Pulvermueller cites Lashley's paper, but he never really addresses metathesis. Instead, he builds his neural model out of "synfire chains", so that in the end the reader finds she has been given a superb 50-year tour of psycholinguistics only to wind up back in 1950, analyzing behavior in terms of neo-Skinnerian stimulus-response chains.

If you are interested in these issues, you should definitely get Pulvermueller's book, if only to understand how cognitive science became locked in this vicious circle of reasoning. To understand Spoonerisms, however, you should look at Loritz' "How the Brain Evolved Language". Unfortunately, Loritz follows the work of Stephen Grossberg, and that requires a really different way of thinking about thought--sort of like the difference between thinking in Newtonian terms and thinking in terms of relativity. I didn't really *get* Loritz (much less Grossberg) until I first read Jeff Hawkins' "On Intelligence". You might want to go this route, too.

Don't be mislead into imagining that just because the book is over 10 years old, that it is obsolete. While newer methods have emerged, the algorithms in the book are still worthy of study by anyone doing research in image analysis or neural nets.

While parts of the book might be too advanced for some readers, it presents a good summary of useful ideas that you can code. Or perhaps start from, if you're doing research.

The book is mostly a first course in numerical methods. Where the authors have provided code examples in C++ and Java. The methods in each topic would typically be found in several other numerical texts. On this basis, the main attraction of the book seems to be the example code.

Readers experienced in C++ or Java coding should be able to write code from scratch to implement methods.

When you have data that is present in some n-dimensional space, you often want to make clusters. The problem is that most methods have a subjective component. What is a cluster is sometimes a matter of definition, within a given method. Clusters can also be used to try to draw up regions of that n-dimensional space. This constitutes a classification of future data. Well, how to do so?

Abe explains an idea that has gained recognition recently. The concept of support vector machines. The label is perhaps a little clumsy. But Abe's book gives a good geometric understanding of current classification ideas and their limitations. And how these can be overcome using support vector machines.

Several variants are explored. Along with a tie-in to neural networks for training. The computations can be intensive for real data. But these days, that is less and less of a limitation.

I was housebound and often bedridden for 15 years with severe Chronic Fatigue Syndrome. I tried many treatment approaches but nothing had any beneficial effect. I began working with Dr. Goldstein in 2001, following the experimental drug protocol set forth in this book. I was on the meds he prescribed for 3 years. During that time, I gradually recovered. I am now off the meds, working fulltime, and living a normal life. I know this approach does not work for everyone, but it has worked for many people with CFS, and I would recommend giving it serious consideration. Dr. Goldstein's hypothesis is that CFS is the result of a malfunction of the nervous system in which the neurons are over-responsive to stimuli. Sounds, smells, bright lights, stress, etc. all over-stimulate the over-reactive nervous system, causing fatigue and adrenal exhaustion. This is certainly consistent with my own experience of the illness. The meds reduce synaptic transmission. Over time, they can restore the nervous system to normal functioning.

Minsky presents interesting new ideas on understanding ourselves. It makes sense that the mind, like the body, may seem simple on the outside but is amazingly complex on the inside.

Like all books on human intelligence also this one is obviously speculation. But it is very helpful speculation if you are interested in the subject. Not all is simple reading especially if you have not been exposed to it before. My own opinion was well advanced by reading this book and I too have agreements and disagreements. Minsky properly suggests that looking at the mind in a too fragmented way will stop us from understanding it. I too see the mind as a complex whole of concepts, processes, and most of all interactive neural resonances. Minsky on the other hand then goes on to fragment the mind's functioning into 20 different capabilities and offers proposals how they work, but enough how they might interact. His perspective is still too fragmente, but then he suggests that uilding those functions into software might create an artifical intelligence.

My own understanding was the most enhanced by his description of how motion and sensation form an expectation loop, meaning that each action has a connected sensation response that leads to another action and so on. This motion/sensation mechanism is built into neural nets as are all other functions of the brain. It allows us to act without abstract thought.

The connection of motion/sensation with abstract thought (often referred to as reason or logical thinking) and emotions requries the contemplation of the human limbic system and neurotransmitters. There I miss quite a few points to be made that Minsky missed.

I also disagree with his expectations into AI. It is this complex resonance between action/sensation, abstract thought and emotion/feelings that creates not only our human experience, but also our intuitive capability. Therefore that will not be emulated by an artificial intelligence mechanism as it lacks our human biological experience that shapes us so strongly. I suggest Steven Pinker's The Blank Slate: The Modern Denial of Human Nature for a better understanding of how we become humans.

My own thoughts of what an AI computer might turn out to be like and how cruel and distorted too rational human beings can be you will find in my own novel Deity. Cruelty is not emotional it is rational.

Overall 'The Emotion Machine' is a great book on the subject and well worth reading if you want to expand your own thinking. The times when a single book was supposed to contain all the dogmatic explanations for our life are gone.

I put this book on the shelf last night about 2/3 done, it will not be finished. I have a back ground in controls engineering and physics simulation, and much of what the author talks about is very similar to these subjects (feedback-control, gain schedualling, non-linear systems, constrained optimization, multi agent interactions etc) yet there is absolutely no use of this body of knowledge. Also the 'style' used in the presented flow diagrams is pain full.

From the word 'machine' in the title I had expected a book with at least a hint of pseudo code or math but got none. I think if you have 'cut metal' on anything related to making a computer do anything that a human can do you will not enjoy this book.

Several other reviews do a good job at describing the content of the book. In contrast I'd like to focus on my reactions to it.

I have read about conciousness before -- including the works of Hofstadler and Penrose ("Emperor's New Mind"). I expected "The Emotion Machine" to be a similar work, exploring the issues of conciousness and the mind in a way that would stimutae further independent thought and ideas about an critical scientific/philosphical issue that mankind and science has been grappling with for ages.

For further personal background, I also have a doctorate in the hard sciences, so I don't think that mist works need to be "dumbed down" for me to understand. Though I do admit there are many smarter and more dedicated people in the world to these issues than myself -- where I'm closer to the "educated layman."

Unfortunately I found this book to be far too textbook-like, much as it was intended not to stand alone but to suppliment lectures and discussions on the topic. I kept getting the "Is this going to be on the test?" feeling and found that this focus did not allow me to shift my focus to the big issues surrounding the mind.

Perhaps the book was intended to be for a different audience than myself, certainly there are other reviewers that found it valuable. For myself however, though given the jacket wording, back cover comments and site summary indicated otherwise, I felt that the book was not valuable to me.

If you're interested in the mind and aren't scared away by some mathematics and scientific method, go instead to the works of Penrose and Hofstadler. This book can be given a pass.

A good book with interesting ideas. However, there was a fair ammount of rehash from his other book (Society of Mind), there was alot of conjecture with not much expereimentally backed up theory, and of course no sourcecode!

This book is officially approved courseware for the CompTIA Network+ exam. It has been approved also by the CSSIA (Center for Systems Security and Information Assurance). The book is very comprehensive. But as all textbooks go it has both strengths and weaknesses. In this review I will try to highlight some of both.

Like most computer industry books, this one comes with a supplemental CD. That CD includes the Certblaster and MeasureUP self-test software and other files the student will find useful.

Each chapter begins with a highlight box that will outline what will be covered in the chapters. Then it includes a letter or story from someone working in the industry. At first these seemed a little cheesy, but as I moved further and further into the book I realized that they were a great source for a feel for the material in the real world, and not just in the classroom or self-study. On the side of the pages are some green bars; these bars highlight specific information you are required to know for the CompTIA exam, and which exam objective they are part of.

One of the greatest strengths is that each chapter ends with a mini glossary of key terms or new terms from that chapter. There is also a complete glossary of all of these terms at the back of the book. Then at each chapter's end there is a series of review questions - multiple choice questions much like those found in self-test software and on the actual exam. Finally, each chapter has a series of hands-on projects. Each project gets progressively more difficult.

The greatest strength of this book is that it has all-encompassing hardware, software, different OS's. The weakness is that it is a lot of material packed very densely into nearly a thousand pages. When I did the school program, we did this book in 3 weeks. For most it was a little overwhelming.

If you want to write the CompTIA Network+ exam, this is a great book to help get you there. Pay special attention to the Novell Netware sections. There were a lot of questions on the exam about those. And Good Luck!

This is a great book for beginners to learn about networking. This was my first networking book for my first networking class, and I will always remember it. I'm reading a lot of bad reviews for this book and I just have to say; this book is not for idiots. If you just "want to pass the Network+" but have no experience, self-study, or interest in networking, then this is not the book for you. If you are interested in networking, and want to learn a lot more, this book does a great job of delivering important information to your brain in an easy to read and enjoyable way. It is a very large book, and contains a LOT of information (yes, some of it is outdated by now, but it's all good to know...if not just for being able to impress your older colleagues/bosses with your knowledge of vampire taps, Banyan Vines, and Thick/Thinnet, then just for a greater understanding of the history behind networking, which will lead you to a broader knowledge of the concepts and enable you to learn new information more easily).

Once again, let me say that this book, despite its title, is NOT good for studying for the Network+. This is a great book to learn the basics of networking from, but when studying for the test you really need to use an Exam Cram 2 and some computer-based practice tests. These practice tests are ESSENTIAL for passing the test. I used TestOut and whatever random ones I could find on the net, but I'm sure they're all pretty equal. If you want to pass the test, get the Exam Cram 2 Network+ book, and install practice test software on your PC.

When you have finished the Exam Cram and can answer every test on the practice test correctly, you will be over-studied for the test, which is the best way to approach any test (especially ones that cost $250). If you're not interested in this stuff enough to spend so much time studying for it, please choose a different career. Leave IT to the people who love it :D

I've had to use this for a Networking+ class and have hated and dreaded having to read it and prep class for it. The writing is vague and sloppy and there are many dubious and wrong statements. Some diagrams do not correspond to the text and/or are confusing. You will not get a good idea of many concepts. It's been very frustrating.
I give it two stars because it is specifically for the Network+ test, but would rate it only one star as a general intro to networking because in addition to the above problems it is full of obsolete and legacy stuff like token ring, coax, NetWare.
The chapter on "Networking with Unix-like OS's" has nothing about networking with them.
My copy will be dumped into the library's free box as soon as the course ends.

Does what it says on the box - it's a good reference, with much detail and useful practical advice for the upcoming network technician or IT support professional. I think its usefulness is fairly limited outside of those fields but as an introduction to computer networking, it will give you plenty of information. Unfortunately the general tone of the book is reflective of its introductory purpose and more experienced readers might find it slightly annoying. If you are a software professional who is more interested in the programming aspects of networking, consider W. Richard Stevens's "TCP/IP Illustrated" series.

I am using this book for my IT class to get my certification for network+.

My classmates and i have discussed the layout of the book and our findings are that while the book contains tons of information that are necessary to get your certification but it was too much to assimilate in a short amount of time.

The problem with this book is that it tries to fill your head with so much text information that by the end of a chapter you are lost because their was too much info to digest. You have to reread the chapters several time in order to try to grasp what the author is trying to teach you in said chapter.


All we get his info over info but we have no clue as to where it is really applied or why we are learning this information at all. It would have been nice to see pictures associated with the concepts that the author was talking about.

As an analogy, i felt like i was trying to get a surgeon's certifcate but all the teaching i got was from a book that had no pictures to show me what to do in order to perform surgeries. How can i operate if i can't make the difference between organs inside the body....

At the end of my readings I did not feel confident at all in my knowledge of this book and was doubtfull i could get my certification on a first try relying on this book information and the way it was presented to me.

This book becomes nothing more than a big memory exercise and if you have a poor memory be prepared to fail your certification.

The only reason we had to use this book in our class is that the company that certifies you is the company that made this book.

I would not recomment this book as an entry level book to this subject as it is too much info to digest in one book.

In my opinion, if you have to read a chapter several times in order to understand what the author is trying to convoy it means to me that the book was not done properly.

This book needs to be revised with the mindset that it has to include more visuals and be an entry level book to certify people for the job.

If you plan on becoming Network+ certfied and you are a huge visual learner stay far way from this book. Get another book that bears the CompTIA seal and check if it is the right book for you.

This book here clearly wasn't the right book for me and my classmates.

Thanks

Plenty of information for make very good projects, unfortunately for my taste, the autor only works with PIC microcontrollers, perhaps a very good book

This book is great for the beginner. Iovine explains the subject matter in a way that makes it exciting and fun. He has a way of getting the exciting vision of building robots across to the reader. The book makes it easy to get parts needed for the projects from the authors web site. I'm really looking forward to his new book, Pic Robotics!!!

I bought this book with the humble desire of creating a simple companion. After I finished, the Creation turned on me and my family and reprogrammed my VCR. It then proceeded to change the message on my answering machine. It somehow convinced my Nissan truck not to allow me inside. I fear for my life. How could I have let it go this far? All I really wanted was a cute little Furby but I ended up with a cyber-monster with dreams of wiping out humans and creating a robotic Utopian Hell.....other than that, the book was pretty cool.

The illustrations and text I found to be very helpful for a project I was working on, but the supplies the book recommends can usually de difficult to find, and can tend to be a bit pricy.
I recommend visiting a local toy store after deciding on a project, and buying toys with the parts you need. Its more fun to make one thing into another anyway.

This was the 3rd book I purchased on pics. The other 2 books - one by Myke Predko (awful), the other by David L. Benson (dissapointing.)

I wish this had been the first. Although not geared specificly towards pics, that was my reason for buying it. I was interested in pics and robotics; so this book was right up my alley.

Admittedly the book has numerous plugs for a company the guy obviously works for, owns, or gets kickbacks from! And he wants you to put out a considerable about of cash from the get go to purchase items he wants you to use in order to follow along with him. However, that doesn't bother me. I never build any projects I see in these type of books. I only use them for learning - I build my own projects.

This book did teach me quite a bit about pics. Which was my goal. He didn't bog you down with the history or innards of pics like other books. Which I am not interested in. The book was a great mixture of hardware and software topics...

I would recommend this book to anyone interested in pics... Subsequently I purchased another book by him simply because I saw his name on it and I wasn't dissapointed! I'm looking forward to other books by John Iovine in the future...