用合成方言语音提升意图识别与槽位填充性能
Aslema at NADI 2026: Data Augmentation for Intent Recognition and Slot Filling
- 用大模型生成突尼斯方言口语，再通过语音克隆合成语音数据
- 混合真实与合成数据训练后，意图识别准确率达86.8%，槽位填充错词率34.7%
- 该方法在官方测试中排名槽位填充第一，适合低资源语言任务研究
我们提出 Aslema 系统，参加 NADI 2026 共享任务5，涵盖意图识别与槽位填充两个子任务。在零样本设置下评估四种通用大模型，并与微调模型对比，结果表明微调始终优于零样本推理。进一步通过大模型生成具有文化背景的突尼斯方言（Tunisian Derja）语句，再经语音克隆生成合成语音，用于数据增强。融合合成数据后，两项任务性能均提升。最终提交系统基于 Qwen3-Omni-30B，使用原始与合成数据混合训练，在开发集上达成 86.8% 意图准确率与 34.7% 错词率（WER）；在官方测试集上，槽位填充排名首位（CoER 59.5），意图识别位列 8 支队伍中的第 4 名（准确率 66.1%）。我们已公开实验脚本，将很快发布合成数据集以支持该领域后续研究。
原文摘要 · Abstract (English)
We present Aslema, our system for NADI 2026 Shared Task 5, which consists of two subtasks: intent recognition and slot filling. We evaluate four omni LLMs in a zero-shot setting and compare them with fine-tuned models. Our results show that fine-tuning consistently outperforms zero-shot inference. We further explore synthetic data augmentation by using an LLM to generate culturally grounded Tunisian Derja utterances, followed by voice cloning to generate synthetic speech. Incorporating this synthetic data improves performance on both tasks. Our final submitted system, based on Qwen3-Omni-30B and trained with a mixture of original and synthetic data, achieves 86.8% intent accuracy and 34.7 WER on the devtest split. On the official test set it ranks 1st in slot filling (59.5 CoER) and 4th among 8 teams in intent recognition (66.1% accuracy). We release our experimental scripts and will soon share the synthetic dataset to support further research in this area.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。