This study explores a pattern emerging from AI research combining artificial intelligence with sound symbolism. Supervised machine learning algorithms, efficient learners of sound symbolism, show a bias towards overpredicting categories perceived as threatening. The study uses XGBoost models on Pokémon names across different languages to test if this bias reflects an adaptation for cautious behavior, as suggested by error management theory.
Publisher
Not specified; study has not undergone peer review.
Published On
Jan 01, 2023
Authors
Alexander Kilpatrick
Tags
artificial intelligence
sound symbolism
machine learning
XGBoost
error management theory
cautious behavior
Pokémon names
Related Publications
Explore these studies to deepen your understanding of the subject.
