Machines

What can SETI learn from AI?

The attempt to construct machine intelligence (AI) may have a great deal to contribute to the search for extra-terrestrial intelligence. First, to the degree that the pursuit of artificial intelligence succeeds, new areas in the domain of potential intelligences are revealed to us. Every successfully-developed mechanical intelligence is, although not extra-terrestrial, to some degree, alien. Second, a significant amount of work in this field is based on lessons learned from biologically-based intelligence observed in nature. AI researchers bring a new perspective to the study of the natural world. Third, theoretical work in computer science and other fields has produced some understanding of the limits of what can be computed and by what means. This avenue of investigation has the potential to rigorously define under what circumstances intelligence may or may not exist, and may ultimately yield many surprising results. Each of these three subtopics is briefly described below.

Artificial Intelligence

John McCarthy, one of the field's founders, defines AI in part as "the science and engineering of making intelligent machines, especially intelligent computer programs." (McCarthy, 2007). AI has been an active area of research for over 50 years, and in that time, much progress has been made. Intelligence may be difficult to define precisely, making it hard to gauge such advances, but in some limited areas of what could commonly be called cognition, machines have already clearly bested their human creators. In many other endeavors, computers, while not matching the human mind, still perform useful "intelligent" functions for the benefit of humanity on a daily basis (Menzies, 2003). Not only is this work useful for the functioning examples of alien intelligence it may produce, but also for the insights into the nature of intelligence which are accumulated by those who pursue it.

Biology and Computation

A strong component of AI research has involved the effort to model natural systems. Although certainly not the first to investigate the phenomenon of intelligent organic systems, AI researchers do bring the unique perspective of those who hope to reverse engineer them. Some of the more significant areas of investigation within this context are:

  • Neural networks
  • Genetic algorithms
  • Artificial life
  • Developmental robotics
  • Cognitive science

Computability Theory

It is a common belief in the field of artificial intelligence that human-level cognition could be implemented on a digital computer given sufficient time and space (Newell and Simon, 1959). (For example, with sufficient resources, it is not difficult to imagine that the brain's physical processes might be simulated in enough detail to reproduce the brain's function.) In addition, a rigorous mathematical branch of computer science called "theory of computation" has established that any reasonable model of computation (including a digital computer) can be simulated by an extremely simple theoretical device called a Turing machine (Sipser, 2005). As a result, by examining the surprisingly diverse ways in which a physical version of a Turing machine might be implemented (Adamatzky, 2001), we may be able to demonstrate that an unexpectedly wide variety of systems are able to support the full depth of human-level intelligence.

Works cited

Adamatzky, A. (2001) Collision-Based Computing. London: Springer-Verlag.

McCarthy, J. (2007) What Is Artificial Intelligence?                                                              http://www-formal.stanford.edu/jmc/whatisai/ [November]

Menzies, T. (2003) 21st-Century AI: Proud, Not Smug. IEEE Intelligent Systems 18(3): 18-24. < http://menzies.us/pdf/03aipride.pdf>

Newell, A., Simon H. (1959) The Simulation of Human Thought. (22 June). http://www.bitsavers.org/pdf/rand/ipl/P-1734_The_Simulation_Of_Human_Thought_Jun59.pdf

Sipser, M. (2005) Introduction to the Theory of Computation. Massachusetts: Thomson Course Technology: 137-159.

Recommended reading

Artificial Intelligence

"AI Topics." 28 July 2008. Association for the Advancement of Artificial Intelligence.
<http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/HomePage>
Accessed 29 May 2009.

Russell, S., Norvig, P. (2002) Artificial Intelligence: A Modern Approach. Upper Saddle River, New Jersey: Prentice Hall.

Biology and Computation

"AI Topics / Neural Networks." 14 December 2008. Association for the Advancement of Artificial Intelligence. < http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/>
Accessed 29 May 2009.

Neural Networks

"AI Topics / Genetic Algorithms." 3 September 2008. Association for the Advancement of Artificial Intelligence.
<http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/>
Accessed 29 May 2009.

Genetic Algorithms

"AI Topics / Artificial Life." 13 December 2008. Association for the Advancement of Artificial Intelligence.
<http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/>
Accessed 29 May 2009.

Artificial Life

"Developmental Robotics." 23 October 2008. Wikipedia.
< http://en.wikipedia.org/w/index.php?title=Developmental_robotics&oldid=247149417 >
Accessed 29 May 2009.

"AI Topics / Developmental Robotics." 14 December 2008. Association for the Advancement of Artificial Intelligence.
<http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/>
Accessed 29 May 2009.

Computability Theory

Sipser, M. (2005) Introduction to the Theory of Computation. Massachusetts: Thomson Course Technology: 137-159.

"Cellular Automaton." 25 April 2009. Wolfram MathWorld.
<http://mathworld.wolfram.com/CellularAutomaton.html>
Accessed 29 May 2009.

"Life Pattern Catalog." 11 February 2001. Paul Callahan.
<http://www.radicaleye.com/lifepage/patterns/contents.html>
Accessed 3 May 2009.

Glossary

Turing machine - a computational device that manipulates symbols on a strip of tape according to a table of rules. Studying the abstract properties of Turing machines yields many insights into computer science and complexity theory.

: Login :
go to project overview go to splash page