# Shun-Ichi Amari **Domain:** Mathematical Neuroscience / Machine Learning / Information Geometry **Doc Type:** Person Node **Maturity:** Foundational ## Definition **Shun-Ichi Amari** is a Japanese mathematical scientist whose early work developed adaptive pattern classifiers, recurrent learning systems and associative-memory architectures. ## 1960s–1970s Context Amari's 1967 adaptive classifier used a probabilistic descent method and explicitly discussed non-global local minima. In 1972 he published learning in recurrent networks and an [[wiki/Associative Memory|associative-memory]] model he later identified with the architecture recognized as the [[wiki/Hopfield Network|Hopfield model]]. ## Sources / Provenance - Shun-Ichi Amari, “A Theory of Adaptive Pattern Classifiers,” 1967. - Shun-Ichi Amari, “Learning Patterns and Pattern Sequences by Self-Organizing Nets of Threshold Elements,” 1972. - [[articles/Modern Artificial Intelligence in the 1970s|Modern Artificial Intelligence in the 1970s]].