Data-driven Natural Language Generation: Paving the Road to Success

Jekaterina Novikova; Ondrej Dusek; Verena Rieser

Data-driven Natural Language Generation: Paving the Road to Success

Jekaterina Novikova, Ondrej Dusek, Verena Rieser

Research output: Contribution to conference › Paper › peer-review

Abstract

We argue that there are currently two major bottlenecks to the commercial use of statistical machine learning approaches for natural language generation (NLG): (a) The lack of reliable automatic evaluation metrics for NLG, and (b) The scarcity of high quality in-domain corpora. We address the first problem by thoroughly analysing current evaluation metrics and motivating the need for a new, more reliable metric. The second problem is addressed by presenting a novel framework for developing and evaluating a high quality corpus for NLG training.

Original language	English
Number of pages	3
Publication status	Published - 30 Jul 2017
Event	First WiNLP Workshop - Vancouver, Canada Duration: 30 Jul 2017 → 30 Jul 2017 http://www.winlp.org/winlp-workshop/

Workshop

Workshop	First WiNLP Workshop
Abbreviated title	WiNLP
Country/Territory	Canada
City	Vancouver
Period	30/07/17 → 30/07/17
Internet address	http://www.winlp.org/winlp-workshop/

Keywords

natural language generation
natural language processing

Access to Document

https://arxiv.org/abs/1706.09433Licence: Other

Cite this

@conference{6209fbff7df041f4adc2a973fe70281d,

title = "Data-driven Natural Language Generation: Paving the Road to Success",

abstract = "We argue that there are currently two major bottlenecks to the commercial use of statistical machine learning approaches for natural language generation (NLG): (a) The lack of reliable automatic evaluation metrics for NLG, and (b) The scarcity of high quality in-domain corpora. We address the first problem by thoroughly analysing current evaluation metrics and motivating the need for a new, more reliable metric. The second problem is addressed by presenting a novel framework for developing and evaluating a high quality corpus for NLG training.",

keywords = "natural language generation, natural language processing",

author = "Jekaterina Novikova and Ondrej Dusek and Verena Rieser",

year = "2017",

month = jul,

day = "30",

language = "English",

note = "First WiNLP Workshop, WiNLP ; Conference date: 30-07-2017 Through 30-07-2017",

url = "http://www.winlp.org/winlp-workshop/",

}

TY - CONF

T1 - Data-driven Natural Language Generation: Paving the Road to Success

AU - Novikova, Jekaterina

AU - Dusek, Ondrej

AU - Rieser, Verena

PY - 2017/7/30

Y1 - 2017/7/30

N2 - We argue that there are currently two major bottlenecks to the commercial use of statistical machine learning approaches for natural language generation (NLG): (a) The lack of reliable automatic evaluation metrics for NLG, and (b) The scarcity of high quality in-domain corpora. We address the first problem by thoroughly analysing current evaluation metrics and motivating the need for a new, more reliable metric. The second problem is addressed by presenting a novel framework for developing and evaluating a high quality corpus for NLG training.

AB - We argue that there are currently two major bottlenecks to the commercial use of statistical machine learning approaches for natural language generation (NLG): (a) The lack of reliable automatic evaluation metrics for NLG, and (b) The scarcity of high quality in-domain corpora. We address the first problem by thoroughly analysing current evaluation metrics and motivating the need for a new, more reliable metric. The second problem is addressed by presenting a novel framework for developing and evaluating a high quality corpus for NLG training.

KW - natural language generation

KW - natural language processing

M3 - Paper

T2 - First WiNLP Workshop

Y2 - 30 July 2017 through 30 July 2017

ER -

Data-driven Natural Language Generation: Paving the Road to Success

Abstract

Workshop

Keywords

Access to Document

Fingerprint

Cite this