@inproceedings{wang-etal-2024-infuserki,
    title = "{I}nfuser{KI}: Enhancing Large Language Models with Knowledge Graphs via Infuser-Guided Knowledge Integration",
    author = "Wang, Fali  and
      Bao, Runxue  and
      Wang, Suhang  and
      Yu, Wenchao  and
      Liu, Yanchi  and
      Cheng, Wei  and
      Chen, Haifeng",
    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/fix-sig-urls/2024.findings-emnlp.209/",
    doi = "10.18653/v1/2024.findings-emnlp.209",
    pages = "3675--3688",
    abstract = "Large Language Models (LLMs) have achieved exceptional capabilities in open generation across various domains, yet they encounter difficulties with tasks that require intensive knowledge. To address these challenges, methods for integrating knowledge have been developed, which augment LLMs with domain-specific knowledge graphs through external modules. These approaches, however, face data inefficiency issues as they necessitate the processing of both known and unknown knowledge for fine-tuning. Thus, our research focuses on a novel problem: efficiently integrating unknown knowledge into LLMs without unnecessary overlap of known knowledge. A risk of introducing new knowledge is the potential forgetting of existing knowledge. To mitigate this risk, we propose the innovative InfuserKI framework. This framework employs transformer internal states to determine when to enrich LLM outputs with additional information, effectively preventing knowledge forgetting. Performance evaluations using the UMLS-2.5k and MetaQA domain knowledge graphs reveal that InfuserKI not only successfully integrates new knowledge but also outperforms state-of-the-art baselines, reducing knowledge forgetting by 9{\%} and 6{\%}, respectively."
}