Towards Interpreting and Mitigating Shortcut Learning Behavior of NLU models
Mengnan Du | Varun Manjunatha | Rajiv Jain | Ruchi Deshpande | Franck Dernoncourt | Jiuxiang Gu | Tong Sun | Xia Hu
Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Recent studies indicate that NLU models are prone to rely on shortcut features for prediction, without achieving true language understanding. As a result, these models fail to generalize to real-world out-of-distribution data. In this work, we show that the words in the NLU training set can be modeled as a long-tailed distribution. There are two findings: 1) NLU models have strong preference for features located at the head of the long-tailed distribution, and 2) Shortcut features are picked up during very early few iterations of the model training. These two observations are further employed to formulate a measurement which can quantify the shortcut degree of each training sample. Based on this shortcut measurement, we propose a shortcut mitigation framework LGTR, to suppress the model from making overconfident predictions for samples with large shortcut degree. Experimental results on three NLU benchmarks demonstrate that our long-tailed distribution explanation accurately reflects the shortcut learning behavior of NLU models. Experimental analysis further indicates that LGTR can improve the generalization accuracy on OOD data, while preserving the accuracy on in-distribution data.
A Semi-Supervised Bayesian Network Model for Microblog Topic Classification
Yan Chen | Zhoujun Li | Liqiang Nie | Xia Hu | Xiangyu Wang | Tat-Seng Chua | Xiaoming Zhang
Proceedings of COLING 2012
Query Segmentation Based on Eigenspace Similarity
Chao Zhang | Nan Sun | Xia Hu | Tingzhu Huang | Tat-Seng Chua
Proceedings of the ACL-IJCNLP 2009 Conference Short Papers
- Tat-Seng Chua 2
- Mengnan Du 1
- Varun Manjunatha 1
- Rajiv Jain 1
- Ruchi Deshpande 1
- show all...