 |
|
|
|
|
||||||
 |
||||||||||
|
Adit Cookbook Pages |
||||||||||
|
||||||||||
|
|
|
|
Growing Up With Lucy - How to Build an Android in Twenty Easy Steps A book review by Mike Griffiths Let me make one point clear, this book does not tell you how to build an android in any number of steps. Despite that, it is still a great (perhaps important) book. What this book might give you is an opportunity to reappraise your perception of modern Artificial Intelligence (AI) theories, techniques and research. It will probably fire you with new ideas and will certainly influence any serious or recreational research and development work you might do in the future whether in robotics, intelligent systems or games. If all of that sounds a little dry then do not worry the style is accessible and often humorous. In fact it is a very good read. If you have not heard of the author, Steve Grand, then let me provide a short and selective biography. Steve Grand is famous (some might say notorious) for being the creator of the computer game "Creatures". This significant software contribution was recognised by an OBE. Despite some honorary fellowships Steve is not an academic but remains a largely self funded and self taught researcher into artificial life and all of it's implications. He rightly refuses to be bound by the limits of what he describes as the compartmentalism of study and draws knowledge and influence from all areas of human knowledge. Steve Grand has an obsession. He wants to understand what he describes as the building blocks of the mind. To develop his ideas and understanding he decided to build a robot and then set about giving it a brain capable of learning and controlling itself. Make that herself. Steve Grand decided to build an android orang-utan and has named her after the earliest known hominid. This new book from Steve Grand makes an effective and, in part, devastating attack on modern AI theory and research. He explains with disarming charm and humour why he finds most of the current theories and research approaches 'suck' or are complete gibberish. He introduces some interesting mind games and experiments that illustrate his points and these will start you, in turn, thinking about the issues and developing theories of your own. You may or may not be swept along in exactly the same direction as the author but you are sure to end up following at least a parallel track. If you are even a distant fan of Occam's razor then you too will favour a simpler, modular theory of the mammalian brain and it's organisation. An organisation that can explain the emergent complexities of vision, muscle control and then perhaps tackle the wonders of intelligence and consciousness. Steve Grand does not have any of the answers as yet but his approach looks likely to lead us in the right direction. So let's look at the book and find out more about Lucy and her development. The book starts with a gentle stroll through binary logic and current work using artificial neural networks. It also covers some of the early design problems that had to be faced in providing any sort of functioning robotic "body" without the resources of a giant Japanese electronics corporation. Of equal importance, the book ponders the need to provide senses and control mechanisms that are suited to the task of working with a brain rather than a digital computer. As we build upon Steve Grand's theories on how vision and hearing is organised within our brains we begin to steadily leave behind any mention of how these may have been applied in program code or electronics. Leaving these aspects of Lucy behind might disappoint those eager to build their own android but we are led upwards and onwards from what has been achieved to what might be achieved with future embodiments of Lucy. The book goes on to explore the mind and ideas about intelligence and that feeling of self awareness we call consciousness. The text also explores the possibility that all of our senses and functions us similar structures and processes to achieve their different functions and ends. You begin to perceive just how it might be possible that something as complex as our own brains could be built from a small set of simple biological building blocks. OK simple is probably the wrong word for these building blocks but the relative simplicity described seems to fit better with our current knowledge of evolution and child development - both key to the creation of each of our individual brains. If Steve Grand's original Lucy is capable of pointing to a banana at least some of the time then this is certainly a magnificent achievement. The book mentions some early work on Lucy Mk II and I am sure that all of this books readers will look forward to the next volume when we can follow the development of Steve Grand and his new baby. In the mean time I have a programming project to re-start as Steve Grand's ideas have made me re-appraise the approach.
Title: Growing Up With Lucy
|
||