@inproceedings{zhang-etal-2026-layer,
    title = "Layer-aware Dual-directional Modulation for Low-resource Machine Translation",
    author = "Zhang, Siqi  and
      Song, Ran  and
      Jiang, Shuting  and
      Huang, Yuxin  and
      Yu, Zhengtao",
    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.1054/",
    pages = "20988--21000",
    ISBN = "979-8-89176-395-1",
    abstract = "Although Large Language Models (LLMs) have achieved remarkable success in Machine Translation (MT), a significant performance gap persists between high-resource and low-resource languages due to imbalanced pre-training data. In this paper, we first investigate the internal mechanisms driving this performance disparity from a layer-wise perspective.We propose a metric termed Activation Disparity ($\Delta R$) to quantify the activation divergence between high- and low-resource MT. Based on this metric, we distinguish between Task-Adaptive Layers (TAL, $\Delta R > 0$) that encode task-specific signals and Legacy-Inert Layers (LIL, $\Delta R < 0$) dominated by pre-trained bias. Leveraging this finding, we propose the \textbf{L}ayer-\textbf{a}ware \textbf{D}ual-directional \textbf{M}odulation (\textbf{LaDM}). Integrated with Low-Rank Adaptation (LoRA), LaDM employs a sparse strategy to bidirectionally modulate optimization dynamics. Specifically, it amplifies contributions from TAL to accelerate feature consolidation while inhibiting LIL to dampen misaligned legacy biases. Extensive experiments on Chinese-to-seven low-resource language translation using Llama-3.1, Qwen2.5, and Gemma-2 demonstrate that LaDM significantly outperforms standard LoRA fine-tuning, achieving an average improvement of 1.73 spBLEU.Code is available at \url{https://github.com/zzssqqq/LaDM}."
}