Multi-domain Neural Network Language Generation for Spoken Dialogue Systems

By Tsung-Hsien Wen, Milica Gasic, Nikola Mrksic, Lina Maria Rojas-Barahona, Pei-hao Su, David Vandyke, Susannah Young. 2016

191 citations recorded by the source corpus.

©2016 Association for Computational Linguistics. Moving from limited-domain natural language generation (NLG) to open domain is difficult because the number of semantic input combinations grows exponentially with the number of domains. Therefore, it is important to leverage existing resources and exploit similarities between domains to facilitate domain adaptation. In this paper, we propose a procedure to train multi-domain, Recurrent Neural Network-based (RNN) language generators via multiple adaptation steps. In this procedure, a model is first trained on counterfeited data synthesised from an out-of-domain dataset, and then fine tuned on a small set of in-domain utterances with a discriminative objective function. Corpus-based evaluation results show that the proposed procedure can achieve competitive performance in terms of BLEU score and slot error rate while significantly reducing the data needed to train generators in new, unseen domains. In subjective testing, human judges confirm that the procedure greatly improves generator performance when only a small amount of data is available in the domain.

Subfields: Artificial Intelligence, Natural Language Processing

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 Natural Language Processing

Citations in the corpus

The E2E Dataset: New Challenges For End-to-End Generation (2017, background): “…development of NLG components in new domains: They jointly learn sentence planning and surface realisation from non-aligned data (Duˇsek and Jurˇc´ıˇcek, 2015; Wen et al., 2015; Mei et al., 2016; Wen et al., 2016; Sharma et al., 2016; Duˇsek and Jurˇc´ıˇcek, 2016a; Lampouras and Vlachos, 2016).”
A Knowledge-Grounded Neural Conversation Model (2017, methodology): “Following (Sordoni et al., 2015; Wen et al., 2016; Li et al., 2016a), we use B LEU automatic evaluation.”
The WebNLG Challenge: Generating Text from RDF Data (2017, methodology): “In (Gardent et al., 2017), we compared a dataset created using the WebNLG process with existing benchmarks in particular, (Wen et al., 2016)’s dataset (RNNLG) which was produced using a similar process.”
Neural Approaches to Conversational AI (2018, neutral): “Picture credit: Wen et al. (2015).”
A Network-based End-to-End Trainable Task-oriented Dialogue System (2016, background): “, 2014) and language generation (Wen et al., 2015b; Wen et al., 2016) modules still rely on supervised learning and therefore need corpora to train on.”
A Survey of Available Corpora For Building Data-Driven Dialogue Systems: The Journal Version (2015, methodology): “…or reinforcement learning framework) using transfer learning (Fabbrizio et al., 2004; Forgues et al., 2014; Serban and Pineau, 2015; Serban et al., 2016; Lowe et al., 2015a; Vandyke et al., 2015; Wen et al., 2016; Gašić et al., 2016; Mo et al., 2016; Genevay and Laroche, 2016; Chen et al., 2016)”
Evaluating the State-of-the-Art of End-to-End Natural Language Generation: The E2E NLG Challenge (2019, methodology): “…domains by reducing annotation effort: They jointly learn sentence planning and surface realisation from non-aligned data, e.g. (Dušek and Jurčíček 2015; Wen et al. 2015b; Mei, Bansal, and Walter 2016; Wen et al. 2016; Sharma et al. 2016a; Dušek and Jurčíček 2016a; Lampouras and Vlachos 2016).”
MultiWOZ - A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modelling (2018, methodology): “What is even worse is that the capability of the initial system introduces additional biases to the collected data, which may result in a mismatch between the training and testing sets (Wen et al., 2016).”
Toward Continual Learning for Conversational Agents (2017, methodology): “There have been prior studies that partly address the data-intensiveness problem outside the neural model (Wen et al., 2016b; Williams et al., 2017; Zhao et al., 2017; Eshghi et al., 2017; Wen et al., 2016a; Zhao et al., 2017).”
Few-shot Natural Language Generation for Task-Oriented Dialog (2020, methodology): “To simulate such a few-shot learning setting, we have developed a new benchmark dataset, called F EW S HOT - WOZ, based on the MultiWOZ (Budzianowski et al., 2018) and Cambridge NLG datasets (Wen et al., 2016a).”
Natural Language Generation in Dialogue using Lexicalized and Delexicalized Data (2016, methodology): “We give a summary of the baseline models we use for performance comparisons in Section 4.7.”
Findings of the E2E NLG Challenge (2018, background): “(Dušek and Jurčı́ček, 2015; Wen et al., 2015b; Mei et al., 2016; Wen et al., 2016; Sharma et al., 2016; Dušek and Jurčı́ček, 2016a; Lampouras and Vlachos, 2016).”
Analysing Data-To-Text Generation Benchmarks (2017, neutral): “(Wen et al., 2016) crowd sourced text for dialogue acts bearing on multiple”
A Survey on Dialogue Systems: Recent Advances and New Frontiers (2017, background): “It was then extended to the multi-domain setting by multiple adaptation steps [96].”
Domain Adaptive Dialog Generation via Meta Learning (2019, background): “Recently, effective domain adaptation has been introduced for natural language generation in dialog systems (Tran and Nguyen, 2018; Wen et al., 2016).”