arXiv cs.CL· Fadel Thior, Ahmer Tabassum, Haania Siddiqui, Hanan Aldarmaki, Thamar Solorio·· 13 小时前AI 评分24
SyRA-CS：基于元学习的合成到真实语码转换 ASR 适配框架
SyRA-CS: Synthetic-to-Real Adaptation for Code-Switched ASR with Meta-Learning
AI 导读
SyRA-CS 是一个基于 Reptile 元学习的语码转换 ASR 合成到真实适配框架，先在各语言对的合成 CS-FLEURS 子集上元训练，再用 1 小时带标注真实目标语音适配。在阿拉伯语、法语、德语、印地语-英语四组语言对、四种架构共 16 组模型对比中，WER 全部优于持续微调，CER 有 15 组更优，XLS-R 的宏观 WER 相对降低 17.7%，其余三个模型降低 5.3-5.8%。
正文 · 原文
Abstract:Code-switched (CS) automatic speech recognition (ASR) remains difficult because real CS speech is scarce and synthetic speech does not fully match it. We investigate whether synthetic CS speech is more useful for learning an adaptation-friendly initialization than for ordinary continual fine-tuning. We propose SyRA-CS, a pair-specific synthetic-to-real adaptation framework based on Reptile meta-learning. For each language pair and pretrained model, we meta-train on the corresponding synthetic CS-FLEURS subset and then adapt using one hour of labeled real target speech. We evaluate four architectural settings: Encoder-CTC, FastConformer-RNN-T, Attention Encoder-Decoder (AED) and Speech-LLM. Experiments cover on Arabic-, French-, German-, and Hindi-English, against synthetic-only transfer, real-only fine-tuning, and continual fine-tuning. SyRA-CS improves over continual fine-tuning on WER in all 16 model-pair comparisons and on CER in 15 of 16, with relative macro WER reductions of 17.7% for XLS-R and 5.3-5.8% for the other three models. The gains hold across the tested models but vary in size across architectures, language pairs and metrics.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.31415 [cs.CL] |
| (or arXiv:2610.31415v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.31415 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Fadel Thior [view email]
[v1]
Thu, 22 Oct 2026 15:17:10 UTC (3,005 KB)
来源：arXiv cs.CL · arxiv.org