Currently, text classification studies mainly focus on training classifiers by using textual input only, or enhancing semantic features by introducing external knowledge (e.g., hand-craft lexicons and domain knowledge). In contrast, some intrinsic statistical features of the corpus, like word frequency and distribution over labels, are not well exploited. Compared with external knowledge, the statistical features are deterministic and naturally compatible with corresponding tasks. In this paper, we propose an Adaptive Gate Network (AGN) to consolidate semantic representation with statistical features selectively. In particular, AGN encodes statistical features through a variational component and merges information via a well-designed valve mechanism. The valve adapts the information flow into the classifier according to the confidence of semantic features in decision making, which can facilitate training a robust classifier and can address the overfitting caused by using statistical features. Extensive experiments on datasets of various scales show that, by incorporating statistical information, AGN can improve the classification performance of CNN, RNN, Transformer, and Bert based models effectively. The experiments also indicate the robustness of AGN against adversarial attacks of manipulating statistical information.
|Title of host publication||Proceedings of the AAAI Conference on Artificial Intelligence|
|Publisher||PKP Publishing Services Network|
|Publication status||Published - 18 May 2021|
|Event||The Thirty-Fifth AAAI Conference on Artificial Intelligence - Online|
Duration: 2 Feb 2021 → 9 Feb 2021
|Name||Proceedings of the AAAI Conference on Artificial Intelligence|
|Conference||The Thirty-Fifth AAAI Conference on Artificial Intelligence|
|Period||2/02/21 → 9/02/21|
Bibliographical noteVol. 35 No. 1: AAAI-21 Technical Tracks 1
Xianming Li’s work has been supported by Ant Group. Zongxi Li’s work has been supported by City University of Hong Kong. Haoran Xie’s work has been supported by the Faculty Research Fund (102041) and the LamWoo Research Fund (LWI20011) at Lingnan University, Hong Kong. Qing Li’s work has been supported by a general research fund from the Hong Kong Research Grants Council (project number:
PolyU 112114/17E). We thank the anonymous reviewers for their careful reading of our manuscript and their insightful comments and suggestions.