Descripción
|
|
---|---|
End-to-end dialog systems are gaining interest due to the recent advances of deep neural networks and the availability of largehuman?human dialog corpora. However, in spite of being of fundamental importance to systematically improve the performanceof this kind of systems, automatic evaluation of the generated dialog utterances is still an unsolved problem. Indeed, most of theproposed objective metrics shown low correlation with human evaluations. In this paper, we evaluate a two-dimensional evalua-tion metric that is designed to operate at sentence level, which considers the syntactic and semantic information carried along theanswers generated by an end-to-end dialog system with respect to a set of references. The proposed metric, when applied to out-puts generated by the systems participating in track 2 of the DSTC-6 challenge, shows a higher correlation with human evalua-tions (up to 12.8% relative improvement at the system level) than the best of the alternative state-of-the-art automatic metricscurrently available | |
Internacional
|
Si |
JCR del ISI
|
Si |
Título de la revista
|
Computer Speech And Language |
ISSN
|
0885-2308 |
Factor de impacto JCR
|
2,116 |
Información de impacto
|
|
Volumen
|
55 |
DOI
|
10.1016/j.csl.2018.12.004 |
Número de revista
|
|
Desde la página
|
200 |
Hasta la página
|
215 |
Mes
|
SIN MES |
Ranking
|