@inproceedings{zhou-etal-2025-difflm,
    title = "{D}iff{LM}: Controllable Synthetic Data Generation via Diffusion Language Models",
    author = "Zhou, Ying  and
      Wang, Xinyao  and
      Niu, Yulei  and
      Shen, Yaojie  and
      Tang, Lexin  and
      Chen, Fan  and
      He, Ben  and
      Sun, Le  and
      Wen, Longyin",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/landing_page/2025.findings-acl.1061/",
    pages = "20638--20658",
    ISBN = "979-8-89176-256-5",
    abstract = "Recent advancements in large language models (LLMs) have significantly enhanced their knowledge and generative capabilities, leading to a surge of interest in leveraging LLMs for high-quality data synthesis. However, synthetic data generation via prompting LLMs remains challenging due to LLMs' limited understanding of target data distributions and the complexity of prompt engineering, especially for structured formatted data. To address these issues, we introduce DiffLM, a controllable data synthesis framework based on variational autoencoder (VAE), which further (1) leverages diffusion models to reserve more information of original distribution and format structure in the learned latent distribution and (2) decouples the learning of target distribution knowledge from the LLM{'}s generative objectives via a plug-and-play latent feature injection module. As we observed significant discrepancies between the VAE{'}s latent representations and the real data distribution, the latent diffusion module is introduced into our framework to learn a fully expressive latent distribution. Evaluations on seven real-world datasets with structured formatted data (i.e., Tabular, Code, and Tool data) demonstrate that DiffLM generates high-quality data, with performance on downstream tasks surpassing that of real data by 2{\%}{--}7{\%} in certain cases. Data and code are available at https://github.com/bytedance/DiffLM."
}