Hugging Face has expanded its Open ASR Leaderboard to include Egyptian and Nigerian English accents. The update comes from a collaboration with Liva AI, a company that collects multimodal interaction data from consumer products. This addition aims to improve representation and trust in automated speech recognition evaluations.
The new data brings the total number of accents on the leaderboard to include Egyptian and Nigerian English. Previously, the list contained Indian-accented English data. The expansion is part of a broader effort to democratize AI through open science, ensuring that models are tested on a wider variety of speech patterns.
Developers and researchers working on speech recognition models will benefit from more accurate benchmarks for diverse dialects. This move helps prevent models from underperforming on non-standard accents. Users building applications for global markets should prioritize testing their systems with these new data sources to ensure reliability across different regions.
