@inproceedings{bohdal-etal-2025-device,
    title = "On-device System of Compositional Multi-tasking in Large Language Models",
    author = "Bohdal, Ondrej  and
      Theodosiadis, Konstantinos  and
      Mpatziakas, Asterios  and
      Filippidis, Dimitrios  and
      Spyrou, Iro  and
      Zonios, Christos  and
      Drosou, Anastasios  and
      Ioannidis, Dimosthenis  and
      Lee, Kyenghun  and
      Moon, Jijoong  and
      Ko, Hyeonmok  and
      Ozay, Mete  and
      Michieli, Umberto",
    editor = "Potdar, Saloni  and
      Rojas-Barahona, Lina  and
      Montella, Sebastien",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track",
    month = nov,
    year = "2025",
    address = "Suzhou (China)",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/name-variant-enfa-fane/2025.emnlp-industry.27/",
    doi = "10.18653/v1/2025.emnlp-industry.27",
    pages = "416--424",
    ISBN = "979-8-89176-333-3",
    abstract = "Large language models (LLMs) are commonly adapted for diverse downstream tasks via parameter-efficient fine-tuning techniques such as Low-Rank Adapters (LoRA). While adapters can be combined to handle multiple tasks separately, standard approaches struggle when targeting the simultaneous execution of complex tasks, such as generating a translated summary from a long conversation. To address this challenge, we propose a novel approach tailored specifically for compositional multi-tasking scenarios involving summarization and translation. Our technique involves adding a learnable projection layer on top of the combined summarization and translation adapters. This design enables effective integration while maintaining efficiency through reduced computational overhead compared to alternative strategies requiring extensive retraining or sequential processing. We demonstrate the practical viability of our method within an on-device environment by developing an Android app capable of executing compositional tasks seamlessly. Experimental results indicate our solution performs well and is fast in both cloud-based and on-device implementations, highlighting the potential benefits of adopting our framework in real-world applications demanding high-speed operation alongside resource constraints."
}