SPOKEN LANGUAGE ACQUISITION BASED ON REINFORCEMENT LEARNING AND WORD UNIT SEGMENTATION

摘要

The process of spoken-language acquisition hasbeen one of the topics of greatest interest to linguistsfor decades. By utilizing modern machine learningtechniques, we simulated this process on computers,which helps to understand it and develop new possibilitiesof applying this concept on intelligent robots,among other things. This paper proposes a newframework for simulating spoken-language acquisitionby combining reinforcement learning and unsupervisedlearning methods. Our experiments alsoshow that a spoken language can be acquired considerablyfaster by identifying potential word segmentsfrom collected ambient sounds in an unsupervisedmanner.