Robot Brains: Circuits and Systems for Conscious Machines [Hardcover]

If you want ideas for building a conscious machine, this gives you one compelling approach. The circuits range from trivial boolean logic to block diagrams, and can't be built directly from this book. If you are an experienced constructor/programmer this is fine for you.
Intriguingly, the author doesn't comment on whether his designs actually achieved his goals. Overall, well worth reading if you are interested in the hardware implmentation of a conscious (not intelligent) machine.

In this book Haikonen lays out a variety of electronic circuits and architectures for constructing robots (and robot simulations), whose lives could then be reasonably characterised as cognitive, possibly even conscious. For those interested in cognitive science and/or artificial neural networks, this book should be an enjoyable, practical and clear introduction to Haikonen's circuits for learning, perception, imagination, kinesthesia, language and other cognitive phenomena.

Haikonen takes a largely synthetic, and mostly rather plausible, approach to machine cognition. He relates his work back well to the cognition and neuroscience of humans and other animals, as well as to the cognitive scientific traditions (e.g. traditional AI, connectionism). He also shows how his architecture may be able to explain various cognitive phenomena, both in a general way - e.g. how a robot might perceive, imagine, integrate perceptual modalities etc, as well as showing how his model may explain some specific cases from the cognitive scientific literature - see e.g. the discussion (pp. 96-97) of modelling the corollary discharge theory synthetically.

Haikonen pitches his explanations very clearly, and makes good and frequent use of diagrams. This reviewer is not an electrical engineer, but could understand just about all of the diagrams and explanations, which is quite an achievement (by Haikonen, not the reviewer :).

As with all science and especially in cognitive science, there are plenty of philosophical commitments in this book, some explicit, others implicit. If the reader is interested in a high-level explication of Haikonen's approach to cognitive machines, I would recommend his 2003 book, "The Cognitive Approach to Conscious Machines".

In summary: Haikonen's book contains plausible and often original proposals for building cognitive robots; and I like the interest Haikonen shows in relating theory to practical matters - an approach seen throughout this stimulating, clearly written, and I would say intellectually quite courageous book. Haikonen tells us what he himself thinks, as well as how cognitive machines might come to think.