Timezone: »

DARE: Disentanglement-Augmented Rationale Extraction
Linan Yue · Qi Liu · Yichao Du · Yanqing An · Li Wang · Enhong Chen


Rationale extraction can be considered as a straightforward method of improving the model explainability, where rationales are a subsequence of the original inputs, and can be extracted to support the prediction results. Existing methods are mainly cascaded with the selector which extracts the rationale tokens, and the predictor which makes the prediction based on selected tokens. Since previous works fail to fully exploit the original input, where the information of non-selected tokens is ignored, in this paper, we propose a Disentanglement-Augmented Rationale Extraction (DARE) method, which encapsulates more information from the input to extract rationales. Specifically, it first disentangles the input into the rationale representations and the non-rationale ones, and then learns more comprehensive rationale representations for extracting by minimizing the mutual information (MI) between the two disentangled representations. Besides, to improve the performance of MI minimization, we develop a new MI estimator by exploring existing MI estimation methods. Extensive experimental results on three real-world datasets and simulation studies clearly validate the effectiveness of our proposed method. Code is released at https://github.com/yuelinan/DARE.

Author Information

Linan Yue (University of Science and Technology of China)
Linan Yue

I am a Ph.D. student at the School of Computer Science and Technology, University of Science and Technology of China (USTC), and a member of the Anhui Province Key Laboratory of Big Data Analysis and Application. My supervisor is Prof Qi Liu. I received my Bachelor degree from HeHai University in July, 2019, and majored in Computer Science and minored in Law. My research interests include Data Mining, Natural Language Processing and Applications of Legal Intelligence.

Qi Liu (" University of Science and Technology of China, China")
Yichao Du (University of Science and Technology of China)
Yanqing An (USTC)
Li Wang (University of Science and Technology of China)
Enhong Chen (University of Science and Technology of China)

More from the Same Authors