COLING 2025main0 citations

Addressing the Training-Inference Discrepancy in Discrete Diffusion for Text Generation

Masaki Asada, Makoto Miwa

Abstract

This study addresses the discrepancy between training and inference in discrete diffusion models for text generation. We propose two novel strategies: (1) a training schema that considers two-step diffusion processes, allowing the model to use its own predicted output as input for subsequent steps during training and (2) a scheduling technique that gradually increases the probability of using self-generated text as training progresses. Experiments conducted on four widely used text generation benchmark datasets demonstrate that both proposed strategies improve the performance of discrete diffusion models in text generation.

BibTeX
@inproceedings{asada-miwa-2025-addressing,
    title = "Addressing the Training-Inference Discrepancy in Discrete Diffusion for Text Generation",
    author = "Asada, Masaki  and
      Miwa, Makoto",
    editor = "Rambow, Owen  and
      Wanner, Leo  and
      Apidianaki, Marianna  and
      Al-Khalifa, Hend  and
      Eugenio, Barbara Di  and
      Schockaert, Steven",
    booktitle = "Proceedings of the 31st International Conference on Computational Linguistics",
    month = jan,
    year = "2025",
    address = "Abu Dhabi, UAE",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.coling-main.477/",
    pages = "7156--7164"
}
Addressing the Training-Inference Discrepancy in Discrete Diffusion for Text Generation · COLING 2025