Falsifiability as the theoretical foundation of Artificial Intelligence

Document Type : علمی - پژوهشی

Abstract

Falsifiability is a criterion for statements for being scientific. This was presented for the first time by Karl Popper - the 20th century philosopher. According to this theory - which is in opposition with the method of induction - a statement is scientific if it can be falsified. New theories will be substituted when the previous theories are falsified. Artificial Intelligence (AI) and learnign machines work in accordance with the falsifiability process. This will become clear by explaining the strategy being implemented in AI. This article, by examining some cases, will show how falsifiability is the theoretical foundation for AI.

Keywords


  1. Allen JF. Bioinformatics and discovery: induction beckons again', Bioessays 2001; 23:104-107.
  2. Alpaydin, Ethem. Introduction to machine learning. MIT press, 2014.
  3. Bishop, Christopher M. Pattern recognition and Machine Learning. 2006.
  4. Duda, Richard O., Peter E. Hart, and David G. Stork. Pattern classification. John Wiley & Sons, 2012.
  5. Gillies Donald. Artificial Intelligence and Scientific Method. Oxford University Press, 1996.
  6. Mitchell, Tom. Machine learning, McGraw Hill, 1997.
  7. Popper, Karl. Conjectures and Refutations, the Growth of Scientific Knowledge. London: Routledge, 1989.
  8. Popper, Karl. The Logic of Scientific Discovery. London: Routledge, 2002.
  9. Popper, Karl. Realism and the Aim of Science. London: Routledge, 1999.
  10. Popper, Karl. Unended Quest, an Intellectual Autobiography. London: Routledge, 1992.
  11. Russell, Stuart Jonathan, Peter Norvig, John F. Canny, Jitendra M. Malik, and Douglas D. Edwards. Artificial intelligence: a modern approach. Vol. 2. Upper Saddle River: Prentice hall, 2003.
  12. Stokes, Geoffrey, Popper, Philosophy, Politics and Scientific Method, Cambridge: Polity Press, 1998.