# Samuel Checkers **Domain:** Machine Learning / Game Playing / AI History **Doc Type:** System Node **Maturity:** Developed **Related:** [[Arthur Samuel]], [[Machine Learning]], [[IBM 701]], [[Artificial Intelligence]] ## Definition **Samuel Checkers** names the evolving family of checkers-playing programs developed by [[Arthur Samuel]] at IBM beginning in the early 1950s. The programs became emblematic of machines improving through play and retained experience. ## Technical Significance Samuel used heuristic evaluation, selective lookahead, and forms of learning to work around the limited memory and compute of systems such as the [[IBM 701]]. The program’s importance was not that checkers itself constituted general intelligence, but that a machine could change future performance using information derived from earlier games. ## Ontology The system is an early bridge among [[Machine Learning]], search, evaluation, memory, and [[Objective Function|objective functions]]. Its closed rules made improvement measurable, while also demonstrating the limits of transferring success from a bounded game to the open world. ## Sources / Provenance - IBM, “The games that helped AI evolve”: https://www.ibm.com/history/early-games