Multi-Domain Joint Semantic Frame Parsing Using Bi-Directional RNN-LSTM

By Dilek Z. Hakkani-Tür, Gökhan Tür, Asli Celikyilmaz, Yun-Nung (Vivian) Chen, Jianfeng Gao, Li Deng, Ye-Yi Wang. 2016

487 citations recorded by the source corpus.

Sequence-to-sequence deep learning has recently emerged as a new paradigm in supervised learning for spoken language understanding. However, most of the previous studies explored this framework for building single domain models for each task, such as slot filling or domain classification, comparing deep learning based approaches with conventional ones like conditional random fields. This paper proposes a holistic multi-domain, multi-task (i.e. slot filling, domain and intent detection) modeling approach to estimate complete semantic frames for all user utterances addressed to a conversational system, demonstrating the distinctive power of deep learning methods, namely bi-directional recurrent neural network (RNN) with long-short term memory (LSTM) cells (RNN-LSTM) to handle such complexity. The contributions of the presented work are three-fold: (i) we propose an RNN-LSTM architecture for joint modeling of slot filling, intent determination, and domain classification; (ii) we build a joint multi-domain model enabling multi-task deep learning where the data from each domain reinforces each other; (iii) we investigate alternative architectures for modeling lexical context in spoken language understanding. In addition to the simplicity of the single model framework, experimental results show the power of such an approach on Microsoft Cortana real user data over alternative methods based on single domain/task deep learning.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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

In Computer Vision and Pattern Recognition

Citations in the corpus

A Progressive Model to Enable Continual Learning for Semantic Slot Filling (2019, background): “They are designed to either train a slot filling model independently (Yao et al., 2014; Peng et al., 2015; Kurata et al., 2016; Hakkani-Tür et al., 2016; Liu and Lane, 2016; Deng et al., 2019; Ray et al., 2019) or jointly with the other intent detection task in SLU (Guo et al., 2014; Liu and Lane,…”
Towards Open Intent Discovery for Conversational Text (2019, background): “[15, 31, 48, 51]) since these are formulated as classification problems, and classify intents into limited, pre-defined categories.”
Dialogue Learning with Human Teaching and Feedback in End-to-End Trainable Task-Oriented Dialogue Systems (2018, background): “We believe that it is because the dialogue state space is too complex for the agent to learn from scratch, as it has to learn a good NLU model in combination with a good policy to complete the task Average Dialogue Turn Size Figure 4 shows the curves for the average turn size of successful dialogues”
A Bi-Model Based RNN Semantic Frame Parsing Model for Intent Detection and Slot Filling (2018, methodology): “It is also possible to use one joint model for intent detection and slot ﬁlling (Guo et al., 2014; Liu and Lane, 2016a,b; Zhang and Wang, 2016; Hakkani-T¨ur et al., 2016).”
End-to-End Task-Completion Neural Dialogue Systems (2017, methodology): “The LU component is implemented with a single LSTM, which performs intent prediction and slot ﬁlling simultaneously (Hakkani-T¨ur et al., 2016; Chen et al., 2016): The LU objective is to maximize the conditional probability of the slots and the intent (cid:126)y given the 1 The source code is…”
A Novel Bi-directional Interrelated Model for Joint Intent Detection and Slot Filling (2019, background): “Hakkani-T¨ur et al. (2016) introduced a RNN-LSTM model where the explicit relationships between the slots and intent are not established.”
Composite Task-Completion Dialogue Policy Learning via Hierarchical Deep Reinforcement Learning (2017, background): “Our composite task-completion dialogue agent consists of four components: (1) an LSTM-based language understanding module (Hakkani-T¨ur et al., 2016; Yao et al., 2014) for identifying user intents and extracting associated slots; (2) a state tracker for tracking the dialogue state; (3) a dialogue…”
Multi-Domain Adversarial Learning for Slot Filling in Spoken Language Understanding (2017, methodology): “Hakkani-Tur et al. [9] proposed a multi-domain SLU model using Bi-directional LSTM.”
Iterative Delexicalization for Improved Spoken Language Understanding (2019, methodology): “Recently, joint intent classification and slot labeling SLU models based on recurrent neural networks has been shown to achieve state-of-the-art performance on benchmark datasets [Hakkani-Tür et al., 2016; Liu and Lane, 2016; Kim et al., 2017a; Goo et al., 2018; Wang et al., 2018].”
Recent advances in deep learning based dialogue systems: a systematic survey (2021, methodology): “Hakkani-Tür et al. (2016) applied a bidirectional RNN-LSTM architecture to jointly perform three tasks.”
A Stack-Propagation Framework with Token-Level Intent Detection for Spoken Language Understanding (2019, methodology): “Considering this strong correlation between the two tasks, some joint models are proposed based on the multi-task learning framework (Zhang and Wang, 2016; Hakkani-T¨ur et al., 2016; Liu and Lane, 2016) and all these models outperform the pipeline models via mutual enhancement between two tasks.”
Neural Approaches to Conversational AI (2018, unknown): “data when creating NLU models for a new domain (Hakkani-Tür et al., 2016; Liu and Lane, 2016).”
Scalable multi-domain dialogue state tracking (2017, methodology): “To tag slots in user utterances, we use multi-domain language understanding (LU) models based on bi-directional recurrent neural networks [15, 16, 17].”
Collaborative Multi-Agent Dialogue Model Training Via Reinforcement Learning (2019, background): “In Hakkani-T¨ur et al. (2016), decoders tag each word in the input sentence with a different slot name and concatenate the intent as a tag to the end-of-sentence token, while in Liu and Lane (2016) the encoder is shared, but the two tasks have separate decoders.”
Out-of-domain Detection based on Generative Adversarial Network (2018, background): “Multi–domain dialog systems (Hakkani-Tur et al., 2016; Jiang et al., 2014; Lee et al., 2013; Ryu et al., 2015; Seon et al., 2014) should detect whether an input request is out–of–domain (OOD) because users do not know the exact coverages of those systems.”
Neural sentence embedding using only in-domain sentences for out-of-domain sentence detection in dialog systems (2017, background): “In contrast, some researchers have worked on supervised sentence embedding particularly for natural language understanding using recurrent neural networks (RNNs) [16, 33, 36] and long short-term memory (LSTM) networks [22, 35, 5].”