Arthur Samuel Proves that Machines Can Learn from Past Errors

In July 1959 Arthur Lee Samuel published "Some Studies in Machine Learning Using the Game of Checkers," IBM Journal of Research and Development 3 (1959) no. 3, 210-29. In this work Samuel demonstrated that machines can learn from past errors — one of the earliest examples of non-numerical computation.

Hook & Norman, Origins of Cyberspace (2002) no. 874.