Sequence-to-Sequence Data Augmentation for Dialogue Language Understanding

Published in

International Conference on Computational Lingu...(2018)

External Links:

Generate Graph

TL;DR

A sequence-to-sequence generation based data augmentation framework that leverages one utterance’s same semantic alternatives in the training data to produce diverse utterances that help to improve the language understanding module.

Abstract

In this paper, we study the problem of data augmentation for language understanding in task-oriented dialogue system. In contrast to previous work which augments an utterance without considering its relation with other utterances, we propose a sequence-to-sequence generation based data augmentation framework that leverages one utterance’s same semantic alternatives in the training data. A novel diversity rank is incorporated into the utterance representation to make the model produce diverse utterances and these diversely augmented utterances help to improve the language understanding module. Experimental results on the Airline Travel Information System dataset and a newly created semantic frame annotation on Stanford Multi-turn, Multi-domain Dialogue Dataset show that our framework achieves significant improvements of 6.38 and 10.04 F-scores respectively when only a training set of hundreds utterances is represented. Case studies also confirm that our method generates diverse utterances.

Authors

Ting Liu

20 papers

Yijia Liu

4 papers

Wanxiang Che

19 papers

References44 items

Neural User Simulation for Corpus-based Policy Optimisation of Spoken Dialogue Systems

Data Augmentation Generative Adversarial Networks

mixup: Beyond Empirical Risk Minimization

Reporting Score Distributions Makes a Difference: Performance Study of LSTM-networks for Sequence Tagging

Key-Value Retrieval Networks for Task-Oriented Dialogue

Sequence-to-Sequence Data Augmentation for Dialogue Language Understanding

Published in

International Conference on Computational Lingu...(2018)

External Links:

Generate Graph

TL;DR

Abstract

Authors

Ting Liu

20 papers

Yijia Liu

4 papers

Wanxiang Che

19 papers

References44 items

Neural User Simulation for Corpus-based Policy Optimisation of Spoken Dialogue Systems

Data Augmentation Generative Adversarial Networks

mixup: Beyond Empirical Risk Minimization

Reporting Score Distributions Makes a Difference: Performance Study of LSTM-networks for Sequence Tagging

Key-Value Retrieval Networks for Task-Oriented Dialogue

Neural Models for Sequence Chunking

OpenNMT: Open-Source Toolkit for Neural Machine Translation

Syntax or semantics? knowledge-guided joint semantic frame parsing

Diverse Beam Search: Decoding Diverse Solutions from Neural Sequence Models

Labeled Data Generation with Encoder-Decoder LSTM for Semantic Slot Filling

End-to-End Memory Networks with Knowledge Carryover for Multi-Turn Spoken Language Understanding

A Sequence-to-Sequence Model for User Simulation in Spoken Dialogue Systems

Data Recombination for Neural Semantic Parsing

Paraphrase Generation from Latent-Variable PCFGs for Semantic Parsing

Leveraging Sentence-level Information with Encoder LSTM for Semantic Slot Filling

Character-level Convolutional Networks for Text Classification

That’s So Annoying!!!: A Lexical and Frame-Semantic Embedding Based Data Augmentation Approach to Automatic Categorization of Annoying Behaviors using #petpeeve Tweets

Effective Approaches to Attention-based Neural Machine Translation

Using Recurrent Neural Networks for Slot Filling in Spoken Language Understanding

Adam: A Method for Stochastic Optimization

Deep Speech: Scaling up end-to-end speech recognition

Spoken language understanding using long short-term memory neural networks

GloVe: Global Vectors for Word Representation

A Systematic Exploration of Diversity in Machine Translation

Recurrent neural networks for language understanding

Paraphrase-Driven Learning for Open Question Answering

Investigation of recurrent-neural-network architectures and learning methods for spoken language understanding

POMDP-Based Statistical Spoken Dialog Systems: A Review

ImageNet classification with deep convolutional neural networks

Hallucinated n-best lists for discriminative language modeling

Syntactic Constraints on Paraphrases Extracted from Parallel Corpora

Cheap and Fast – But is it Good? Evaluating Non-Expert Annotations for Natural Language Tasks

Error simulation for training statistical dialogue systems

Agenda-Based User Simulation for Bootstrapping a POMDP Dialogue System

A probabilistic framework for dialog simulation and optimal strategy learning

Paraphrasing with Bilingual Parallel Corpora

Extracting Paraphrases from a Parallel Corpus

Probabilistic simulation of human-machine dialogues

A speech understanding system based on statistical representation of semantics

Evaluation of Spoken Language Systems: the ATIS Domain

GENERATIVE ADVERSARIAL NETS

Learning user simulations for information state update dialogue systems

Human-computer dialogue simulation using hidden Markov models

Corpus-based dialogue simulation for automatic strategy learning and evaluation

Field of Study

Computer Science

Venue Information

Name

International Conference on Computational Linguistics

Type

conference

URL

https://www.aclweb.org/anthology/venues/coling/

Alternate Names

Int Conf Comput Linguistics
COLING