aiqbench analysis
This research demonstrates the efficacy of fine-tuning large language models over zero-shot inference for specific NLP tasks like intent recognition and slot filling. The study further validates synthetic data augmentation strategies, utilizing LLMs to generate culturally grounded Tunisian Derja utterances and voice cloning to enhance model performance on low-resource dialects. By releasing scripts and datasets, the work provides a valuable resource for advancing few-shot learning techniques in diverse linguistic contexts.
Aslema at NADI 2026: Augmentation through Fewshot for SLU
arXiv:2608.18689v1 Announce Type: cross Abstract: 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.
Abstract: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.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2608.18689 [cs.CL]
(or arXiv:2608.18689v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2608.18689 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 247 words
Related coverage
NADI 2026: The Second Multidialectal Arabic Speech Processing Shared Task
arXiv · Sep 24, 2026
EAGT: Echocardiography Augmentation for Generalisability and Transferability
arXiv · Jul 29, 2026
Structural Preservation Governs Data Augmentation in Deep Learning-Based Laser Speckle Material Classification
arXiv · Jul 28, 2026
Enhancing LLMs for Identifying and Prioritizing Important Medical Jargons from Electronic Health Record Notes Utilizing Data Augmentation: A Comparative Study
arXiv · Jul 23, 2026