WAXAL-NET: 针对19种非洲语言的微调边缘ASR
WAXAL-NET: Finetuned Edge ASR Across 19 African Languages
AI总结：
本研究评估了紧凑型领域专用ASR模型在WAXAL语料库的19种非洲语言会话语音上是否优于大规模多语言基础模型，通过微调边缘模型实现了宏平均WER从64.9%降至38.0%，模型大小缩小3-40倍，证实领域专业化主导规模效应。
摘要：
我们评估了紧凑型领域专用ASR模型是否能在WAXAL语料库的19种非洲语言会话语音上优于大规模多语言基础模型。微调后的边缘模型实现了宏平均词错误率（WER）38.0%，而最佳零样本基线为64.9%，使用小3-40倍的模型降低了26.9个百分点。结果证实，对于自发的非洲语音，领域专业化主导规模效应。跨域评估显示，微调模型在分布外（OOD）语音上恢复了可用性能，而零样本模型在测试域与其预训练分布匹配时重新获得优势。一项涵盖所有调查语言的分布式母语者审计产生了基于语言学的错误分类，表明CTC和自回归架构在不同语系中表现不同。我们进一步表明，对于音节文字语言，仅WER会错误表示性能，其中CER/WER比率显示字符级准确率远高于标题WER所暗示的。最后，为促进未来的非洲ASR研究，我们发布了所有模型权重、微调和评估脚本，以及涵盖全部19种语言的清洗后的WAXAL子集。
英文摘要：
We evaluate whether compact domain-specialized ASR models can outperform massively multilingual foundation models for conversational African speech across 19 languages in the WAXAL corpus. Fine-tuned edge models achieve a macro-averaged WER of $38.0\%$ compared to $64.9\%$ for the best zero-shot baseline, a $26.9$ percentage-point reduction using models $3-40\times$ smaller. Results confirm that domain specialization dominates scale for spontaneous African speech. Cross-domain evaluation shows that fine-tuned models recover usable performance on out-of-distribution (OOD) speech, while zero-shot models regain an advantage when the test domain matches their pretraining distribution. A distributed native-speaker audit across all surveyed languages produces a linguistically-grounded error taxonomy, showing that CTC and autoregressive architectures behave differently across language families. We further show that WER alone misrepresents performance for syllabary-script languages where CER/WER ratios reveal substantially higher character-level accuracy than headline WER suggests. Finally, to contribute to future African ASR research, we release all model weights, fine-tuning and evaluation scripts, and a cleaned WAXAL subset covering all $19$ languages.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
