Evaluating AI supported ASR performance: the role of semantic density within the DLM

Read full text
Authors: Motoki Saito, Hanno Müller, Annette Gerstenberg
Cited by: 0

Abstract

AI-supported Automatic Speech Recognition (ASR) performance varies depending on age-related factors (e.g., Verhaeghen, 2003).Greater age → Bigger vocabulary → More meanings compete → More difficult to map word form onto meaning (i.e., greater semantic uncertainty)?A poor ASR performance may be due to increased uncertainty in the mapping between form and meaning.We test:whether semantic uncertainty predicts ASR performance.whether computational models with vocabularies representing different ages predict ASR performance more precisely.

Fields of Study & Subjects

Speech Recognition and Synthesis Phonetics and Phonology Research Neurobiology of Language and Bilingualism

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
26 February 2026
Language
EN