Talking to myself: self-dialogues as data for conversational agents

Joachim Fainberg; Ben Krause; Mihai Dobre; Marco Damonte; Emmanuel Kahembwe; Daniel Duma; Bonnie Webber; Federico Fancellu

Talking to myself: self-dialogues as data for conversational agents

Computation and Language 2018-09-20 v2 Artificial Intelligence

Authors: Joachim Fainberg , Ben Krause , Mihai Dobre , Marco Damonte , Emmanuel Kahembwe , Daniel Duma , Bonnie Webber , Federico Fancellu

View on arXiv ↗ PDF ↗

Abstract

Conversational agents are gaining popularity with the increasing ubiquity of smart devices. However, training agents in a data driven manner is challenging due to a lack of suitable corpora. This paper presents a novel method for gathering topical, unstructured conversational data in an efficient way: self-dialogues through crowd-sourcing. Alongside this paper, we include a corpus of 3.6 million words across 23 topics. We argue the utility of the corpus by comparing self-dialogues with standard two-party conversations as well as data from other corpora.

Keywords

dialogue systems corpus natural language processing

Cite

@article{arxiv.1809.06641,
  title  = {Talking to myself: self-dialogues as data for conversational agents},
  author = {Joachim Fainberg and Ben Krause and Mihai Dobre and Marco Damonte and Emmanuel Kahembwe and Daniel Duma and Bonnie Webber and Federico Fancellu},
  journal= {arXiv preprint arXiv:1809.06641},
  year   = {2018}
}

Comments

5 pages, 5 pages appendix, 2 figures

Talking to myself: self-dialogues as data for conversational agents

Abstract

Keywords

Cite

Comments

Related papers