Code: https://github.com/apple-yinhan/FTL4LALM

Dataset: https://huggingface.co/datasets/apple121/MMAU-Pro-Ctrl 

Separation Model: https://huggingface.co/apple121/FTL4LALM 

If you find this work useful, please cite: 

@inproceedings{yin2026focusthen,

  title     = {Focus Then Listen: An Empirical Study of Plug-and-Play Audio Enhancer for Noise-Robust Large Audio Language Models},

  author    = {Han Yin and Yang Xiao and Younghoo Kwon and Ting Dang and Jung-Woo Choi},

  booktitle = {ICML 2026 Workshop on Machine Learning for Audio (Learning to Listen)},

  year      = {2026},

  url       = {https://mlforaudioworkshop.github.io/}

}