The Archimedean trap: Why traditional reinforcement learning will probably not yield AGI

Journal of Artificial General Intelligence 11 (1):70-85 (2020)
  Copy   BIBTEX


After generalizing the Archimedean property of real numbers in such a way as to make it adaptable to non-numeric structures, we demonstrate that the real numbers cannot be used to accurately measure non-Archimedean structures. We argue that, since an agent with Artificial General Intelligence (AGI) should have no problem engaging in tasks that inherently involve non-Archimedean rewards, and since traditional reinforcement learning rewards are real numbers, therefore traditional reinforcement learning probably will not lead to AGI. We indicate two possible ways traditional reinforcement learning could be altered to remove this roadblock.



External links

  • This entry has no external links. Add one.
Setup an account with your affiliations in order to access resources via your University's proxy server

Through your library

Similar books and articles

Measurement without archimedean axioms.Louis Narens - 1974 - Philosophy of Science 41 (4):374-393.
Some determinants of rigidity in discrimination-reversal learning.Arnold H. Buss - 1952 - Journal of Experimental Psychology 44 (3):222.
The growth of learning during non-differential reinforcement.Allen D. Calvin - 1953 - Journal of Experimental Psychology 46 (4):248.
Determinants of the effects of vicarious reinforcement.Albert R. Marston - 1966 - Journal of Experimental Psychology 71 (4):550.


Added to PP

11,616 (#309)

6 months
195 (#14,910)

Historical graph of downloads
How can I increase my downloads?

Author's Profile

Samuel Allen Alexander
Ohio State University (PhD)

Citations of this work

Add more citations

References found in this work

Utilitarianism.J. S. Mill - 1861 - Oxford University Press UK. Edited by Roger Crisp.
Non-standard Analysis.Gert Heinz Müller - 2016 - Princeton University Press.
Non-Archimedean Probability.Vieri Benci, Leon Horsten & Sylvia Wenmackers - 2013 - Milan Journal of Mathematics 81 (1):121-151.

View all 15 references / Add more references