aiqbench analysis
FSA-GRPO introduces a reinforcement learning-based post-training method that enables auditory LLMs to effectively utilize few-shot demonstrations for low-resource tasks like children's speech recognition. By leveraging high-resource adult ASR data, this approach improves general adaptation capabilities without requiring in-domain training sets, offering a practical solution for scenarios where specific domain data is unavailable. The technique demonstrates superior performance over direct tuning on out-of-domain data, highlighting its potential for enhancing audio understanding and speech translation in data-scarce environments.
FSA-GRPO: Teaching Auditory LLMs to Use Few-shot Demonstrations
arXiv:2606.02615v1 Announce Type: cross Abstract: Few-shot prompting provides an effective way to adapt auditory large language models to low-resource tasks such as children's speech recognition. However, most auditory large language models are not explicitly trained to perform inference in this demonstration-conditioned format, limiting the extent to which they can benefit from few-shot prompting. To address this limitation, we introduce Few-Shot Aware GRPO (FSA-GRPO), an RL-based post-training recipe that uses a specially designed reward to encourage the model to leverage few-shot demonstrations, thereby strengthening its few-shot adaptation ability. Notably, training with only high-resource adult ASR data improves the model's general few-shot adaptation ability, yielding gains not only in children's speech recognition but also in speech translation and audio understanding. We further study data selection and auxiliary reward weighting to identify an effective training recipe. Our experiments show that when in-domain data are unavailable or cannot be used for training, FSA-GRPO is more effective than direct tuning on related out-of-domain data.
Abstract:Few-shot prompting provides an effective way to adapt auditory large language models to low-resource tasks such as children's speech recognition. However, most auditory large language models are not explicitly trained to perform inference in this demonstration-conditioned format, limiting the extent to which they can benefit from few-shot prompting. To address this limitation, we introduce Few-Shot Aware GRPO (FSA-GRPO), an RL-based post-training recipe that uses a specially designed reward to encourage the model to leverage few-shot demonstrations, thereby strengthening its few-shot adaptation ability. Notably, training with only high-resource adult ASR data improves the model's general few-shot adaptation ability, yielding gains not only in children's speech recognition but also in speech translation and audio understanding. We further study data selection and auxiliary reward weighting to identify an effective training recipe. Our experiments show that when in-domain data are unavailable or cannot be used for training, FSA-GRPO is more effective than direct tuning on related out-of-domain data.
Audio and Speech Processing (eess.AS); Artificial Intelligence (cs.AI); Sound (cs.SD)
arXiv:2606.02615 [eess.AS]
(or arXiv:2606.02615v1 [eess.AS] for this version)
https://doi.org/10.48550/arXiv.2606.02615 Focus to learn more arXiv-issued DOI via DataCite
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 · 253 words
Related coverage
FSA-GRPO: Teaching Auditory LLMs to Use Few-Shot Demonstrations
arXiv · Sep 1, 2026
AFDBench: A Reasoning-First AI Scientist for NationalWeather Service Forecast Discussions
arXiv · Aug 27, 2026
Diverse reasoning traces teach LLMs to make better decisions
Amazon Science · May 26, 2026
Bypassing inference bottlenecks: Accelerating complex AI search with Retrieve-for-Train
Google Research · Sep 15, 2026