@inproceedings{li-etal-2026-lightweight,
    title = "Lightweight Haar Wavelet Subband Pruning for {LLM}s",
    author = "Li, Jiang  and
      Cao, Pengfei  and
      Zhou, Chenxi  and
      Lan, Tian  and
      Su, Xiangdong  and
      Liu, Kang  and
      Zhao, Jun  and
      Gao, Guanglai",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.798/",
    pages = "16242--16259",
    ISBN = "979-8-89176-395-1",
    abstract = "Large language models (LLMs) reach state-of-the-art performance across many NLP tasks, but their large parameter counts introduce heavy computational and memory overhead, which complicates deployment in resource-constrained settings. Pruning is a standard compression strategy that induces sparsity to lower these costs. However, most pruning methods for LLMs depend on calibration data and expensive weight updates, which limits practical scalability. To address these limitations, we introduce \textbf{H}aar \textbf{W}avelet \textbf{S}ubband \textbf{P}runing (), a post-training framework that requires no calibration data and no weight updates. applies a two-dimensional Haar wavelet transform to each weight matrix and decomposes it into four frequency subbands. It then assigns a uniform sparsity ratio to all subbands so that both low- and high-frequency components are retained in a balanced manner. Our theoretical analysis shows that the subband design of provides a deterministic per-subband retention guarantee, which helps mitigate the potential bias of global magnitude pruning toward dominant frequency components. Experiments on the LLaMA, OPT and Qwen model families show that achieves competitive accuracy relative to strong pruning baselines while substantially reducing pruning time. Compared with magnitude pruning, which serves as a simple calibration-free baseline, generally achieves better downstream performance across a wide range of sparsity levels and model scales."
}