Dialog State Tracking: A Neural Reading Comprehension Approach

By Shuyang Gao, Abhishek Sethi, Sanchit Agarwal, Tagyoung Chung, Dilek Z. Hakkani-Tür. 2019

169 citations recorded by the source corpus.

Dialog state tracking is used to estimate the current belief state of a dialog given all the preceding conversation. Machine reading comprehension, on the other hand, focuses on building systems that read passages of text and answer questions that require some understanding of passages. We formulate dialog state tracking as a reading comprehension task to answer the question what is the state of the current dialog? after reading conversational context. In contrast to traditional state tracking methods where the dialog state is often predicted as a distribution over a closed set of all the possible slot values within an ontology, our method uses a simple attention-based neural network to point to the slot values within the conversation. Experiments on MultiWOZ-2.0 cross-domain dialog dataset show that our simple system can obtain similar accuracies compared to the previous more complex methods. By exploiting recent advances in contextual word embeddings, adding a model that explicitly tracks whether a slot value should be carried over to the next turn, and combining our method with a traditional joint state tracking method that relies on closed set vocabulary, we can obtain a joint-goal accuracy of 47.33% on the standard test split, exceeding current state-of-the-art by 11.75%**.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Continual Prompt Tuning for Dialog State Tracking (2022, background): “To alleviate the reliance on the ontology and improve generalization to unseen values, some work extract values from a dialog context (Xu and Hu, 2018; Gao et al., 2019) while others generate values directly to handle situations where values are missing from the context (Wu et al.”
Multi-domain Dialogue State Tracking as Dynamic Knowledge Graph Enhanced Question Answering (2019, methodology): “Early models of dialogue state tracking (Thomson & Young, 2010; Wang & Lemon, 2013; Henderson et al., 2014) rely on handcrafted features to extract utterance semantics, and then use these features to predict dialogue states. Recently Mrkšić et al. (2017) propose to use convolutional neural network t”
Schema-Guided Multi-Domain Dialogue State Tracking with Graph Attention Neural Networks (2020, methodology): “For example, Neural Reading DST model (Gao et al. 2019), which utilizes an attention-based neural network to point to the slot values in the utterances; HyST (Goel, Paul, and Hakkani-T¨ur 2019), a hybrid joint state tracking model learning the optimal method for each slot type, and TRADE (Wu et al.…”
Recent Advances in Natural Language Processing via Large Pre-trained Language Models: A Survey (2021, background): “[50] formulated the dialogue state tracking as a QA problem.”
Event Extraction as Machine Reading Comprehension (2020, background): “Our work also relates to works connecting MRC and other tasks, such as relation extraction (Levy et al., 2017; Li et al., 2019b), semantic role labeling (FitzGerald et al., 2018), named entity recognition (Li et al., 2019a), and others (Wu et al., 2019; Gao et al., 2019).”
MinTL: Minimalist Transfer Learning for Task-Oriented Dialogue Systems (2020, methodology): “MinTL -based systems consistently outperform the DAMD (Zhang et al., 2019b), MD-Sequicity (Lei et al., 2018) (Eric et al., 2019) 40.20 38.00 HyST (Goel et al., 2019) † 44.24 - SUMBT (Lee et al., 2019) † 46.65 - TOD-BERT (Wu et al., 2020) - 48.00 DST-Picklist (Zhang et al., 2019a) - 53.30 SST (Chen e”
TripPy: A Triple Copy Strategy for Value Independent Neural Dialog State Tracking (2020, methodology): “Chao and Lane (2019) and Gao et al. (2019) solely rely on span-prediction, but their performance lacks behind methods that at least partially rely on a pre-deﬁned list of candidate values.”
Find or Classify? Dual Strategy for Slot-Value Predictions on Multi-Domain Dialog State Tracking (2019, methodology): “In addition, DST-Span and DS-DST both achieve better performance when compared with DSTreader.”
MultiWOZ 2.1: Multi-Domain Dialogue State Corrections and State Tracking Baselines (2019, methodology): “r with a copy mechanism (Wu et al., 2019). The DST Reader is a newly proposedmodelthatframesstatetrackingasareadingcomprehensionproblem,learningtoextractslot valuesas spans from the dialogue history (Gao et al., 2019). The HyST is another new model which combines a hierarchical encoder fixed vocabul”
MultiWOZ 2. (2019, background): “The DST Reader is a newly proposed model that frames state tracking as a reading comprehension problem, learning to extract slot values as spans from the dialogue history (Gao et al., 2019).”
TOD-BERT: Pre-trained Natural Language Understanding for Task-Oriented Dialogue (2020, methodology): “We also list several well-known dialogue state trackers as reference, including DSTReader (Gao et al., 2019), HyST (Goel et al., 2019), TRADE (Wu et al., 2019), and ZS-DST (Rastogi et al., 2019).”
Preview, Attend and Review: Schema-Aware Curriculum Learning for Multi-Domain Dialogue State Tracking (2021, background): “Due to the language variety of multi-turn dialogs, the concepts of slots and values are often indirectly expressed in the conversation (such as co-references, ellipsis, and diverse appearances), which are a major bottleneck for improving DST performance (Gao et al., 2019; Hu et al., 2020).”
Leveraging Slot Descriptions for Zero-Shot Cross-Domain Dialogue StateTracking (2021, methodology): “Rastogi et al. (2020) incorporated slot descriptions for facilitating cross domain DST, while Gao et al. (2019, 2020) formulated DST as a question answering problem by casting a slot name into questions.”
Data Augmentation for Copy-Mechanism in Dialogue State Tracking (2020, unknown): “[Gao et al. , 2019] used the pretrained language model BERT [Devlin et al. , 2018] and copy-mechanism to predict explicitly expressed values.”