@inproceedings{na-etal-2024-enhancing,
    title = "Enhancing Large Language Model Based Sequential Recommender Systems with Pseudo Labels Reconstruction",
    author = "Na, Hyunsoo  and
      Gang, Minseok  and
      Ko, Youngrok  and
      Seol, Jinseok  and
      Lee, Sang-goo",
    editor = "Al-Onaizan, Yaser  and
      Bansal, Mohit  and
      Chen, Yun-Nung",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024",
    month = nov,
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest_wac_2008/2024.findings-emnlp.423/",
    doi = "10.18653/v1/2024.findings-emnlp.423",
    pages = "7213--7222",
    abstract = "Large language models (LLMs) are utilized in various studies, and they also demonstrate a potential to function independently as a recommendation model. Nevertheless, training sequences and text labels modifies LLMs' pre-trained weights, diminishing their inherent strength in constructing and comprehending natural language sentences. In this study, we propose a reconstruction-based LLM recommendation model (ReLRec) that harnesses the feature extraction capability of LLMs, while preserving LLMs' sentence generation abilities. We reconstruct the user and item pseudo-labels generated from user reviews, while training on sequential data, aiming to exploit the key features of both users and items. Experimental results demonstrate the efficacy of label reconstruction in sequential recommendation tasks."
}