@inproceedings{ragazzi-etal-2024-token,
    title = "What Are You Token About? Differentiable Perturbed Top-$k$ Token Selection for Scientific Document Summarization",
    author = "Ragazzi, Luca  and
      Italiani, Paolo  and
      Moro, Gianluca  and
      Panni, Mattia",
    editor = "Ku, Lun-Wei  and
      Martins, Andre  and
      Srikumar, Vivek",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2024",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2024.findings-acl.561/",
    doi = "10.18653/v1/2024.findings-acl.561",
    pages = "9427--9440",
    abstract = "Scientific document summarization aims to condense complex and long articles in both technical and plain-language terms to facilitate the accessibility and dissemination of scientific findings. Existing datasets suffer from a deficiency in source heterogeneity, as their data predominantly stem from a single common resource, hindering effective model training and generalizability. First, we introduce SciLay, a novel dataset that includes documents from multiple natural science journals with expert-authored technical and lay summaries. Second, we propose PrunePert, a new transformer-based model that incorporates a differentiable perturbed top-$k$ encoder layer to prune irrelevant tokens in end-to-end learning. Experimental results show that our model achieves a nearly 2x speed-up compared to a state-of-the-art linear transformer, remaining comparable in effectiveness. Additional examinations underscore the importance of employing a training dataset that includes different sources to enhance the generalizability of the models. Code is available at https://github.com/disi-unibo-nlp/sci-lay."
}