Research Papers 论文研究 8d ago Updated 7d ago 更新于 7天前 43

Comparative Analysis of Multilingual Pre-trained Models for Nepali Automatic Speech Recognition 尼泊尔自动语音识别多语言预训练模型比较分析

First controlled benchmark comparing six multilingual pretrained ASR models on Nepali under identical fine-tuning protocols, using the OpenSLR SLR54 corpus (~165 hours) Whisper-Large-v3-Turbo (14.76% WER) and IndicWav2Vec (14.89% WER) achieve near-identical top performance despite a 9x parameter gap and 40x pretraining data gap CTC-based decoders run up to 29x faster than autoregressive Whisper models at equivalent accuracy, strongly favoring CTC for latency-sensitive deployments Massively multi 首次对6种多语言预训练模型在尼泊尔语ASR任务上进行统一协议下的公平对比,填补了标准化基准的空白 Whisper-Large-v3-Turbo(14.76% WER)与IndicWav2Vec(14.89% WER)性能并列第一,尽管前者参数多9倍、预训练数据多40倍 CTC架构解码器在相同精度下比自回归Whisper快29倍,颠覆了延迟敏感场景的部署偏好 大规模多语言预训练(MMS-1B)在域外数据(FLEURS)上退化最小(+12.55 pp),证明规模换取的是鲁棒性而非峰值精度

55
Hot 热度
72
Quality 质量
58
Impact 影响力

Analysis 深度分析

TL;DR

  • First controlled benchmark comparing six multilingual pretrained ASR models on Nepali under identical fine-tuning protocols, using the OpenSLR SLR54 corpus (~165 hours)
  • Whisper-Large-v3-Turbo (14.76% WER) and IndicWav2Vec (14.89% WER) achieve near-identical top performance despite a 9x parameter gap and 40x pretraining data gap
  • CTC-based decoders run up to 29x faster than autoregressive Whisper models at equivalent accuracy, strongly favoring CTC for latency-sensitive deployments
  • Massively multilingual MMS-1B shows the smallest out-of-domain degradation on FLEURS (+12.55 pp), demonstrating that scale improves robustness over peak in-domain accuracy
  • Language-family proximity in pretraining can effectively substitute for raw model scale when targeting in-domain Nepali speech recognition

Why It Matters

This benchmark fills a critical gap for low-resource and underrepresented language communities by providing the first standardized, multi-model, efficiency-aware reference for Nepali ASR. For AI practitioners building speech systems for Indic languages, it offers actionable guidance on trading off accuracy, latency, and out-of-domain robustness. The finding that family-proximity can replace scale has broader implications for resource-constrained multilingual ASR development.

Technical Details

  • Six pretrained models evaluated: XLSR-53, IndicWav2Vec, MMS-1B, Whisper-Medium, Whisper-Large-v3-Turbo, and Conformer-Hi, spanning CTC self-supervised, autoregressive encoder-decoder, and hybrid Conformer-CTC architectures
  • Fine-tuning conducted on OpenSLR SLR54 Nepali corpus (~165 hours) with identical preprocessing, data splits, optimizer, and family-matched learning-rate schedules across all models
  • Evaluation metrics include Word Error Rate (WER), Character Error Rate (CER), and Real-Time Factor (RTF) across three independent test sets: OpenSLR, FLEURS, and Common Voice
  • Key quantitative results: Whisper-Large-v3-Turbo and IndicWav2Vec both achieve ~14.8% WER; CTC decoders achieve up to 29x speedup over autoregressive Whisper at comparable accuracy; MMS-1B shows +12.55 pp out-of-domain degradation on FLEURS, the smallest among all models tested

Industry Insight

  • For production Nepali ASR systems, CTC-based architectures (e.g., XLSR-53, IndicWav2Vec) should be the default choice when latency is a constraint, as they match top accuracy at a fraction of the compute cost of large autoregressive models
  • When targeting Indic languages with limited in-domain data, prioritizing models pretrained on language-family-proximate data (e.g., IndicWav2Vec for Nepali) can be more effective than scaling up parameter count alone
  • For applications requiring strong out-of-domain generalization across diverse speech conditions, massively multilingual models like MMS-1B offer superior robustness, making them suitable for deployment in noisy or variable real-world environments

TL;DR

  • 首次对6种多语言预训练模型在尼泊尔语ASR任务上进行统一协议下的公平对比,填补了标准化基准的空白
  • Whisper-Large-v3-Turbo(14.76% WER)与IndicWav2Vec(14.89% WER)性能并列第一,尽管前者参数多9倍、预训练数据多40倍
  • CTC架构解码器在相同精度下比自回归Whisper快29倍,颠覆了延迟敏感场景的部署偏好
  • 大规模多语言预训练(MMS-1B)在域外数据(FLEURS)上退化最小(+12.55 pp),证明规模换取的是鲁棒性而非峰值精度

为什么值得看

本文为低资源语言ASR系统选型提供了首个标准化、多模型、效率感知的参考基准,对开发尼泊尔语及类似语言语音产品的团队具有直接指导价值。研究揭示了语言家族亲缘性可替代原始规模的核心发现,为资源受限场景下的模型选择提供了新的决策维度。

技术解析

  • 模型范围:涵盖6种预训练模型(XLSR-53、IndicWav2Vec、MMS-1B、Whisper-Medium、Whisper-Large-v3-Turbo、Conformer-Hi),覆盖CTC自监督、自回归编码器-解码器和混合Conformer-CTC三种架构类型
  • 训练协议:使用OpenSLR SLR54尼泊尔语语料库(约165小时),采用统一的预处理、数据划分、优化器和家族匹配的学习率调度策略,确保对比公平性
  • 评估体系:在三个独立测试集(OpenSLR、FLEURS、Common Voice)上评估WER、CER和RTF,兼顾域内精度与域外泛化能力
  • 核心发现:IndicWav2Vec凭借语言家族相近性(印欧语系)实现与Whisper-Large-v3-Turbo相当的性能,证明预训练数据的语言亲缘性可弥补规模差距;CTC解码器在相同准确率下实现29倍推理加速

行业启示

  • 模型选型策略:对于延迟敏感的低资源语言部署场景,应优先考虑CTC架构而非盲目追求大参数自回归模型,可在精度损失可控的前提下获得数量级的推理加速
  • 预训练数据价值重估:语言家族亲缘性可能是比预训练数据规模更关键的因子,建议在开发南亚/东南亚语言ASR时优先选用同语系预训练模型
  • 鲁棒性与精度权衡:大规模多语言模型(如MMS-1B)虽非峰值精度最优,但在域外数据上表现最稳健,适合对泛化能力要求高的生产环境

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Speech 语音 Benchmark 基准测试 Fine-tuning 微调 Dataset 数据集 Research 科学研究