Behind Deep Blue : Building the Computer that Defeated the World Chess Champion which gives an interesting overview of how the machine developed over time, as well as what it's like to be in graduate school / industrial research. It also gives the reader an appreciation for how interesting AI applications can "devolve" into engineering.
The Mind's I, by Hofstadter and Dennett. A very interesting (and fun) exploration of "philosophy of mind" ideas.
How the Mind Works, Steven Pinker. Scholarly, yet (somehow) also easy to read. Smart discussions about a whole host of interesting topics, like the evolution of intelligence, what are emotions good for, and so on.
The Age of Spiritual Machines, Ray Kurzweil. A very optimistic view of what the progress toward building intelligent machines is likely to be. Paints a picture in which there’s not an us (people)/them (machines) but rather a scenario in which we’ve so augmented ourselves by machines that it’s hard to tell the difference. Raises a lot of practical and ethical questions about what will happen if we succeed at this.
Robot: Mere Machine to Transcendent Mind. Hans Moravec. Projections for the future of AI and robotics by an accomplished roboticist.
The Sciences of the Artificial, Herbert Simon. Describes sciences that study teleologically motivated systems, e.g., the mammalian heart or a cell phone company or a robot, that were designed or were evolved to accomplish some goal (as opposed, say, to a volcano, that happens to do something but that’s not why it exists). Old, but very insightful, from one of the first AI researchers and the winner of a Nobel prize in economics.
Vehicles: Experiments in Synthetic Psychology, Valentino Braitenberg . It's short, fun to read, and it raises interesting issues about what it means to be intelligent/alive/ have feelings etc. From an Amazon review: “These imaginative thought experiments are the inventions of one of the world's eminent brain researchers. They are "vehicles," a series of hypothetical, self-operating machines that exhibit increasingly intricate if not always successful or civilized "behavior." Each of the vehicles in the series incorporates the essential features of all the earlier models and along the way they come to embody aggression, love, logic, manifestations of foresight, concept formation, creative thinking, personality, and free will. In a section of extensive biological notes, Braitenberg locates many elements of his fantasy in current brain research. ”
Blondie 24: Playing at the Edge of AI, David Fogel. It is it is somewhat technical but presented in a more generally appealing manner than most monographs. From Amazon reviews: “The writing is surprisingly engaging, coming from a software researcher; even readers with little interest in checkers will follow Fogel's many game analyses with rising interest as his neural networks increase in prowess.” “Dr. Fogel did serious work in a true scientific manner, and the book is just fine, but potential readers should be aware that the playing strength of Blondie24 is greatly exaggerated. In testing against other checker playing computer programs Blondie24 is completely undistinguished and cannot play an even match with any sort of serious checker program. This points to the fallacy of the implied premise of the book. In fully deterministic games, neural networks to date have been a remarkable failure. In probabalistic games such as Backgammon, on the other hand, they have been an incredible success, as the world's top Backgammon programs are based on neural nets.”
Speech and Language Processing, Jurafsky and Martin.
Reinforcement Learning, Sutton and Barto