@inproceedings{2c88fb68c4db437db6ff9b4847524547,
title = "Learning credible deep neural networks with rationale regularization",
abstract = "Recent explainability related studies have shown that state-of-the-art DNNs do not always adopt correct evidences to make decisions. It not only hampers their generalization but also makes them less likely to be trusted by end-users. In pursuit of developing more credible DNNs, in this paper we propose CREX, which encourages DNN models to focus more on evidences that actually matter for the task at hand, and to avoid overfitting to data-dependent bias and artifacts. Specifically, CREX regularizes the training process of DNNs with rationales, i.e., a subset of features highlighted by domain experts as justifications for predictions, to enforce DNNs to generate local explanations that conform with expert rationales. Even when rationales are not available, CREX still could be useful by requiring the generated explanations to be sparse. Experimental results on two text classification datasets demonstrate the increased credibility of DNNs trained with CREX. Comprehensive analysis further shows that while CREX does not always improve prediction accuracy on the held-out test set, it significantly increases DNN accuracy on new and previously unseen data beyond test set, highlighting the advantage of the increased credibility.",
keywords = "Credibility, Deep neural network, Expert rationales, Explainability",
author = "Mengnan Du and Ninghao Liu and Fan Yang and Xia Hu",
note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 19th IEEE International Conference on Data Mining, ICDM 2019 ; Conference date: 08-11-2019 Through 11-11-2019",
year = "2019",
month = nov,
doi = "10.1109/ICDM.2019.00025",
language = "English (US)",
series = "Proceedings - IEEE International Conference on Data Mining, ICDM",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "150--159",
editor = "Jianyong Wang and Kyuseok Shim and Xindong Wu",
booktitle = "Proceedings - 19th IEEE International Conference on Data Mining, ICDM 2019",
address = "United States",
}