Is it Deja Vu All Over Again? Part II
Welcome back to the ongoing story of our current drama!
SUMMARY
Cybernetics and connections between Cybernetics and the brain. Some additional history related to neurophysiology and neuroscience. The McCulloch-Pitts neuron. Can the machine mimic the brain? The concept of the “intelligent machine”. Disciplines involved in the early Cybernetics movement.
Cybernetics – Control and Communication Theory, Whether in the Machine or the Animal
Before we leave the war and post-war years of the 1940s, discussed in our first episode of the evolution of AI (the would-be monster), you might be interested to learn more about the work related to human or animal biology, in the form of neurophysiology and neuroscience. These were active fields of inquiry which were emerging at the same time. Indeed the structure and basic operation of the biological neuron and synaptic connections had been described, and it was natural that a connection would be found with the new work being done in computing and control.
We recall that the war effort had focused the attention of Dr. Norbert Wiener on the possibility of automatic control of artillery and the design and development of the soon-to-be-realized electronic digital computer.
In the process of contemplating problems of this kind, Wiener had turned his attention to how the brain itself might integrate the sensory and motor actions of the body in performing simple tasks (even the picking up of a pencil – an example of a problem in control). He and his co-worker realized very quickly that a feedback mechanism between sensory and control functions must be at work, They predicted that if the feedback between sensory inputs and neural outputs in the brain were excessive, the arm, in attempting to carry out the task of picking up a pencil could go into a state of uncontrolled oscillation. On the other hand, if the nervous system failed to transmit the required signals successfully, the action might become impossible, resulting in a state of “ataxia”.
In fact, upon inquiry to the medical community, both conditions were known, and in the former case, the condition was known as “purpose tremor” associated with injury to the cerebellum, and in the latter case, it could show up as “tabes dorsalis” which can be a result of the action of syphilis on the nervous system.
To Wiener, the mathematical study of control and communication opened a new window on “that part of physiology which concerns not solely the elementary processes of nerves and synapses, but the performance of the nervous system as an integrated whole”.
When Cybernetics formally received its name in 1947, it was introduced as a name for “the entire field of control and communication theory, whether in the machine or in the animal”. The similarities and possible connections between the operation of the biological brain and nervous system, and the operation of the (computing) machine had been duly recognized.
Communications and Control in the Human Nervous System – the McCulloch-Pitts Neuron and the Turing Machine
It is interesting to note that this kind of work, relating to communications and control in the human nervous system, was active as early as 1942. In that year, the Macy Foundation provided support for a meeting in New York devoted to problems related to the nervous system. You might recognize some of the names who were present at that meeting. They included, in addition to Dr. Wiener, Dr. Warren McCulloch of the Medical School of the University of Illinois. Dr. McCulloch had already been in touch with Dr. Wiener concerning the possible organization of the cortex (outer layers) of the brain.
It would be McCulloch and another mathematician, Walter Pitts, who would collaborate on early work related to “the union of nerve fibers by synapses into systems with specific properties”. The McCulloch-Pitts neuron (1943) would ultimately be embraced by some members of the AI community. To some extent, this model is still in use today. However, today it should probably be thought of as an oversimplified neural model.
Turing (late 1930s) had conceived of a kind of computing machine, known as a Turing machine, which can be thought of as a kind of “universal computer”, capable of computing any computable function. The brain can do the same, using networks of neurons in communication with each other. In addition, neurons can receive specific inputs from the sensory organs and are capable of responding with physiological control responses such a moving an arm or a leg. And this is not to mention what the mind itself can do.
In any event, without belaboring these points, it was natural to ask at that time whether a computing machine might be able to mimic the actions of the brain, and therefore eventually become an “intelligent machine” (whatever that might encompass). Along these lines, it is also easy to see the connection to future work in robotics which was on the horizon.
Early Disciplines Involved in the Cybernetics Movement
We want to make the point that, at this time, the people who were participating in the Cybernetics movement represented a wide range of disciplines interested in the brain-machine connection. These disciplines included but were not necessarily limited to engineering, physics, mathematics, psychology, physiology, anthropology, sociology, and neuroscience, and all were represented in formal meetings held at that time.
And now the story starts to come together, leading to the official recognition and efforts to discover the “intelligent machine”. And with that we end Part II and leave you in eager anticipation as our story of Deja Vu continues to unfold!
What’s Next?
Stay tuned for Part III in our ongoing saga that will lead us ever closer to the present state-of-the-art in AI and all the interesting questions and answers associated with that.
Some specific quotations here are from the book “Cybernetics” by Norbert Wiener