Pre-trained language models have been successful on text classification tasks, but are prone to learning spurious correlations from biased datasets, and are thus vulnerable when making inferences in a new domain. Prior work reveals such spurious patterns via post-hoc explanation algorithms which compute the importance of input features. Further, the model is regularized to align the importance scores with human knowledge, so that the unintended model behaviors are eliminated. However, such a regularization technique lacks flexibility and coverage, since only importance scores towards a pre-defined list of features are adjusted, while more complex human knowledge such as feature interaction and pattern generalization can hardly be incorporated. In this work, we propose to refine a learned language model for a target domain by collecting human-provided compositional explanations regarding observed biases. By parsing these explanations into executable logic rules, the human-specified refinement advice from a small set of explanations can be generalized to more training examples. We additionally introduce a regularization term allowing adjustments for both importance and interaction of features to better rectify model behavior. We demonstrate the effectiveness of the proposed approach on two text classification tasks by showing improved performance in target domain as well as improved model fairness after refinement.
Cross-Refine: Improving Natural Language Explanation Generation by Learning in Tandem
Technische Universität Berlin（柏林工业大学）；Saarland Informatics Campus（萨尔兰信息学园区）；Centre for European Research in Trusted AI (CERTAIN)（欧洲可信人工智能研究中心）；German Research Center for Artificial Intelligence (DFKI)（德国人工智能研究中心）
Latent Logic Tree Extraction for Event Sequence Explanation from LLMs
School of Computer Science and Engineering, Nanyang Technological University（南洋理工大学计算机科学与工程学院）；School of Data Science, The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）数据科学学院）；Skywork AI（Skywork AI公司）
Teaching Machine Comprehension with Compositional Explanations
Department of Computer Science, University of Southern California（南加州大学计算机科学系）；Information Science Institute, University of Southern California（南加州大学信息科学研究所）
Uncertainty-Aware Explainable Recommendation with Large Language Models
Chengdu University of Information Technology（成都信息工程大学）；Tunghai University（东海大学）；Leshan Normal University（乐山师范学院）；University at Buffalo, SUNY（纽约州立大学布法罗分校）；Meta；Purdue University（普渡大学）；University at Albany, SUNY（纽约州立大学奥尔巴尼分校）
Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images
Allen Institute for Artificial Intelligence（艾伦人工智能研究所）；The Hebrew University of Jerusalem（耶路撒冷希伯来大学）；Ben Gurion University of the Negev（内盖夫本古里安大学）；University of Washington（华盛顿大学）
