Freund & Schapire's AdaBoost explicitly adapted Littlestone & Warmuth's multiplicative weight-update rule, and won the 2003 Gödel Prize for it

machine-learning-theoryboostingonline-learninghistory-of-computer-scienceadaboost

Yoav Freund and Robert Schapire's 1995/97 paper, "A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting" — the paper that introduces AdaBoost — names its debt in print: "we show that the multiplicative weight-update rule of Littlestone and Warmuth [10] can be adapted to this model yielding bounds that are slightly weaker in some cases." AdaBoost's core mechanism — reweighting training examples after each round so subsequent weak learners focus on what prior rounds got wrong, then combining the weak learners into a strong one — is a direct adaptation of the 1989 Weighted Majority Algorithm's multiplicative-update rule, moved from the online-expert-prediction setting into the boosting setting.

This is a real, cited inheritance, not a rediscovery under the same name: unlike RA-RAG's silent 2025 reinvention of "weighted majority voting", Freund and Schapire named their source and adapted its proof technique rather than its vocabulary. ACM SIGACT awarded the paper the 2003 Gödel Prize, calling it "a permanent contribution to science even beyond computer science" — an institutional record of the paper's standing, not evidence for the mechanism claim above, which rests entirely on the 1995 primary.