Evolving Intelligence

In the last chapter we endowed a machine, our crane, with rudimentary intelligence. In its fourth generation of evolution, it could sense several types of changes in its environment and make appropriate compensations in its behavior to meet its goal, namely transferring boxes from one conveyor to another. 

The boxes can be of different sizes and contain different products. Some of these products are light and delicate, some are heavy and rugged. The heavy boxes with rugged products in them need to be tightly held by the grippers, lest they slip out and fall. That same grip applied to 

As before, the spacing and orientation of the boxes on the input conveyor belt vary some are properly aligned some are turned at odd angles. Sometimes there is a long time interval before the arrival of the next box sometimes several of them are jammed together. 

Our first evolutionary step is to add new senses to our crane/sorter. Several of these needs can be met with simple all-or-none-type sensors. Our output conveyors must be fitted with sensors that detect not only whether they are running, but whether they are running in the proper direction. A sensor to measure high temperatures would serve as fire detectors. A simple voltage sensor would measure whether power is available to the crane. 

Dealing with various sizes and orientations of boxes is more difficult. Our fourth-generation crane used a television camera above the input conveyor belt, coupled to an analysis program that found the edges and shape, such that it could recognize a box, its orientation, and its location. To do this, it had to consult an internal memory that defined what a box looked like in a television picture. 

---

Moreover, behaviors frequently arise that appear to involve some form of intelligent division of red and blue forces to deal with local firestorms and skirmishes, particularly those forces whose personalities have been evolved (via a genetic algorithm sec below) to perform a specific mission, It is important to point out that such behaviors are not hard-wired but rather an emergent property of a decentralized and nonlinear local dynamics, A small sampling of behaviors is shown below. 

The method is different in nature from the techniques discussed elsewhere in this book because, despite the fact that the cells that comprise a CA model evolve, they do not learn. The lack of a learning mechanism places the method at the boundaries of Artificial Intelligence, but the algorithm still has features in common with AI methods and offers an intriguing way to solve scientific problems. 

With more links in the chain, brains have more opportunities to make adjustments—there are more knobs and controls with which to fine-tune the activities of neurons. Current theories of learning and memory identify the plasticity—adjustability—of synapses as the mechanism by which brains acquire and store information. Intelligence evolves from the continual tweaking of chemical transmission in the brain. 

Robert Jastrow, the head of NASA s Goddard Institute for Space Studies, called this the most powerful evidence for the existence of God ever to come out of science. Other amazing parameters abound. If all of the stars in the universe were heavier than three solar masses, they would live for only about 500 million years, and life would not have time to evolve beyond primitive bacteria. Stephen Hawking has estimated that if the rate of the universe s expansion one second after the Big Bang had been smaller by even one part in a hundred thousand million million, the universe would have recollapsed.The universe must live for billions of years to permit time for intelligent life to evolve. On the other hand, the universe might have expanded so rapidly that protons and electrons never united to make hydrogen atoms. 

Today s Internet evolved from the tiny seed planted by Ts ai Lun. Both paper and the Internet break the barriers of time and distance, and permit unprecedented growth and opportunity. In the next decade, communities formed by ideas will be as strong as those formed by geography. The Internet will dissolve away nations as we know them today. Humanity becomes a single hive mind, with a group intelligence, as geography becomes putty in the hands of the Internet sculptor. 

Highly evolved four-dimensional creatures would have extraordinarily developed nervous systems due in part to the increased number of possible neuronal synapses. Therefore the beings would be super-intelligent. I would expect new and higher senses to be present. For example, propioception, or pressure sensing, would probably not occur perpendicular to a 2-D creature s plane of existence. Similarly, four-dimensional creatures would have senses and sensory receptors that would make little sense to us. 

Becerra, O., Mexican cartels evolve in the face of cross-border crackdown . Jane s Intelligence Review, November 2003. 

---