@inproceedings{jang-etal-2021-kw,
    title = "{KW}-{ATTN}: Knowledge Infused Attention for Accurate and Interpretable Text Classification",
    author = "Jang, Hyeju  and
      Bang, Seojin  and
      Xiao, Wen  and
      Carenini, Giuseppe  and
      Ng, Raymond  and
      Lee, Young ji",
    editor = "Agirre, Eneko  and
      Apidianaki, Marianna  and
      Vuli{\'c}, Ivan",
    booktitle = "Proceedings of Deep Learning Inside Out (DeeLIO): The 2nd Workshop on Knowledge Extraction and Integration for Deep Learning Architectures",
    month = jun,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/Author-page-Marten-During-lu/2021.deelio-1.10/",
    doi = "10.18653/v1/2021.deelio-1.10",
    pages = "96--107",
    abstract = "Text classification has wide-ranging applications in various domains. While neural network approaches have drastically advanced performance in text classification, they tend to be powered by a large amount of training data, and interpretability is often an issue. As a step towards better accuracy and interpretability especially on small data, in this paper we present a new knowledge-infused attention mechanism, called KW-ATTN (KnoWledge-infused ATTentioN) to incorporate high-level concepts from external knowledge bases into Neural Network models. We show that KW-ATTN outperforms baseline models using only words as well as other approaches using concepts by classification accuracy, which indicates that high-level concepts help model prediction. Furthermore, crowdsourced human evaluation suggests that additional concept information helps interpretability of the model."
}