The Role of Syntactic Planning in Compositional Image Captioning
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Standard
The Role of Syntactic Planning in Compositional Image Captioning. / Bugliarello, Emanuele; Elliott, Desmond.
Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. Online : Association for Computational Linguistics, 2021. p. 593–607.Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Harvard
of the Association for Computational Linguistics, 21/04/2021. https://doi.org/10.18653/v1/2021.eacl-main.48
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - The Role of Syntactic Planning in Compositional Image Captioning
AU - Bugliarello, Emanuele
AU - Elliott, Desmond
N1 - Conference code: 16
PY - 2021/4
Y1 - 2021/4
N2 - Image captioning has focused on generalizing to images drawn from the same distribution as the training set, and not to the more challenging problem of generalizing to different distributions of images. Recently, Nikolaus et al. (2019) introduced a dataset to assess compositional generalization in image captioning, where models are evaluated on their ability to describe images with unseen adjective–noun and noun–verb compositions. In this work, we investigate different methods to improve compositional generalization by planning the syntactic structure of a caption. Our experiments show that jointly modeling tokens and syntactic tags enhances generalization in both RNN- and Transformer-based models, while also improving performance on standard metrics.
AB - Image captioning has focused on generalizing to images drawn from the same distribution as the training set, and not to the more challenging problem of generalizing to different distributions of images. Recently, Nikolaus et al. (2019) introduced a dataset to assess compositional generalization in image captioning, where models are evaluated on their ability to describe images with unseen adjective–noun and noun–verb compositions. In this work, we investigate different methods to improve compositional generalization by planning the syntactic structure of a caption. Our experiments show that jointly modeling tokens and syntactic tags enhances generalization in both RNN- and Transformer-based models, while also improving performance on standard metrics.
U2 - 10.18653/v1/2021.eacl-main.48
DO - 10.18653/v1/2021.eacl-main.48
M3 - Article in proceedings
SP - 593
EP - 607
BT - Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume
PB - Association for Computational Linguistics
CY - Online
T2 - The 16th Conference of the European Chapter<br/> of the Association for Computational Linguistics
Y2 - 21 April 2021 through 23 April 2021
ER -
ID: 275339891