LogicalFactChecker: Leveraging Logical Operations for Fact Checking with Graph Module Network

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 567–582 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 567 Efficient Dialogue State Tracking by Selectively Overwriting Memory Sungdong Kim Sohee Yang Gyuwan Kim Sang-Woo Lee Clova AI, NAVER Corp. {sungdong.kim, sh.yang, gyuwan.kim, sang.woo.lee}@navercorp.com Abstract Recent works in dialogue state tracking (DST) focus on an open vocabulary-based setting to resolve scalability and generalization issues of the predefined ontology-based approaches. However, they are inefficient in that they pre- dict the dialogue state at every turn from scratch. Here, we consider dialogue state as an explicit fixed-sized memory and propose a selectively overwriting mechanism for more efficient DST. This mechanism consists of two steps: (1) predicting state operation on each of the memory slots, and (2) overwrit- ing the memory with new values, of which only a few are generated according to the predicted state operations. Our method de- composes DST into two sub-tasks and guides the decoder to focus only on one of the tasks, thus reducing the burden of the decoder. This enhances the effectiveness of training and DST performance. Our SOM-DST (Se- lectively Overwriting Memory for Dialogue State Tracking) model achieves state-of-the- art joint goal accuracy with 51.72% in Mul- tiWOZ 2.0 and 53.01% in MultiWOZ 2.1 in an open vocabulary-based DST setting. In ad- dition, we analyze the accuracy gaps between the current and the ground truth-given situa- tions and suggest that it is a promising direc- tion to improve state operation prediction to boost the DST performance.1 1 Introduction Building robust task-oriented dialogue systems has gained increasing popularity in both the research and industry communities (Chen et al., 2017). Di- alogue state tracking (DST), one of the essential tasks in task-oriented dialogue systems (Zhong et al., 2018), is keeping track of user goals or in- tentions throughout a dialogue in the form of a set of slot-value pairs, i.e., dialogue state. Because the 1The code is available at github.com/clovaai/som-dst. Figure 1: An example of how SOM-DST performs dia- logue state tracking at a specific dialogue turn (in this case, fifth). The shaded part is the input to the model, and “Dialogue State at turn 5” at the right-bottom part is the output of the model. Here, UPDATE operation needs to be performed on the 10th and 11th slot. DST at this turn is challenging since the model requires reason- ing over the long-past conversation. However, SOM- DST can still robustly perform DST because the pre- vious dialogue state is directly utilized like a memory. next dialogue system action is selected based on the current dialogue state, an accurate prediction of the dialogue state has significant importance. Traditional neural DST approaches assume that all candidate slot-value pairs are given in advance, i.e., they perform predefined ontology-based DST (Mrkˇsi´c et al., 2017; Zhong et al., 2018; Nouri and Hosseini-Asl, 2018; Lee et al., 2019). Most previ- ous works that take this approach perform DST by scoring all possible slot-value pairs in the ontology and …

Content truncated to 500 words