Presented at InterSymp-2019 Conference, International Institute for Advanced Studies in Systems Research and Cybernetics, July 1, 2019.
This paper reviews recent advances in automated computer-based learning capabilities. It briefly describes and examines the strengths and weaknesses of the five principal algorithmic approaches to machine-learning, namely: connectionism; evolutionism; Bayesianism; analogism; and, symbolism. While each of these approaches can demonstrate some degree of learning, a learning capability that is comparable with human learning is still in its infancy and will likely require the combination of multiple algorithmic approaches. However, the current state reached in machine-learning suggests that Artificial General Intelligence and even Artificial Superintelligence may indeed be eventually feasible.
© Jens Pohl 2019
Number of Pages