In “Intelligence Without Reason”, Rodney Brooks discusses the current issues with traditional Artificial Intelligence approach to solving problems and proposes an alternative approach based upon his research. Brooks mentions that computers and though are two categories that together define Artificial Intelligence as a discipline. He proposes that the state of computer architecture has been a strong influence on our models of thought. Intelligence within biological systems is completely different. Recent work in behavior-based Artificial Intelligence has produced new models of intelligence that are much closer in spirit to biological systems. The non-Von Neumann computational models they use share many characteristics with biological computations.

A reasonable characterization of the general field of AI is that it is intended to make computers do things, that when done by people, are described as having indicated intelligence. Recently, there has been a movement to study intelligence from the bottom up, concentrating on physical systems (e.g., mobile robotics), situated in the world, autonomously carrying out tasks of various sorts. There is a requirement that intelligence be reactive to dynamic aspects of the environment, that a mobile robot operate on time scales similar to those of animals and humans, and that intelligence be able to generate robust behavior in the face of uncertain sensors, an unpredictable environment, and a changing world. There are questions on how the architecture of our computers influence our choice of problems on which to work, our models of thought, and our algorithms, and how the problems on which we work, our models of thought, and our algorithm choice puts pressure on the development of architectures of our computers.

For intelligence, Turing posed a question, “Can machines think?” To tease our an acceptable meaning for this question he presented what has come to be known as the “Turing Test”, where a person communicates in English over a teletype with either another person or a computer. The goal is to guess whether it is a person or a computer at the other end. Another area was development of software for AI logic, where McCarthy developed the Lisp programming language. McCarthy’s Lisp was much clean and simpler than the alternatives. It made processing lists of information and recursive tree searches trivial to program. Search procedures now became even easier and more convenient to include in AI programs. AI research in this area lead to the area of Cybernetics. Much of the work in Cybernetics really was aimed at understanding animals and intelligence. Animals were modeled as machines and from those models, it was hoped to glean how the animals changed their behavior through learning, and how that lead to better adaptation to the environment for the whole organism.

One flaw with traditional AI is the “knowledge is everything” approach where the system will know everything my reading its own knowledge base instead of interacting and learning from its environment. There was research done in the late fifties and sixties involving linear threshold devices to emulate neurons. This lead to look at AI from a biological perspective. We have our own introspection to tell us how our minds work and our own observations to tell us how the behavior of other people and animals works. The common view in AI is that there is a central storage system which links together the information about concepts, goals, desires etc. There is a tendency to believe that the knowledge is stored in a way that is independent from the way or circumstances in which it was acquired, which has been shown to not be entirely true.

The behavior based robot uses a different approach to solving problems by using the world as its model, continuously referring to its sensors rather than to an internal world model. Instead of having representations of individual entities in the world, the system has representations in terms of its relationship of the entities to the robot. These relationships are both spatial and functional. It is hard to identify the seat of intelligence within any system, as intelligence is produced by the interactions of many components. Intelligence can only be determined by the total behavior of the system and how that behavior appears in relation to the environment. As the complexity of the world increases, and the complexity of processing to deal with the world arises, we will see the same evidence of though and consciousness in our systems as we see in people other than ourselves now. Thought and consciousness will not be programmed in but will emerge.

Reference:

  1. Brooks, Rodney. “Intelligence Without Reason.” <http://people.csail.mit.edu/brooks/papers/AIM-1293.pdf>