Papers with systematic error differences
Language technology practitioners as language managers: arbitrating data bias and predictive bias in ASR (2022.lrec-1)
Copied to clipboard
| Challenge: | despite natural language variation, automatic speech recognition systems perform worse on non-standardised and marginalised language varieties. |
| Approach: | They propose a re-framing of language resources as (public) infrastructure for speech communities . authors propose rethinking of algorithms to address the origins and harms of bias . |
| Outcome: | The proposed approach aims to understand the origins and harms of algorithmic bias and how it can be mitigated. |