TY - GEN
T1 - Latent-Space Data Augmentation for Visually-Grounded Language Understanding
AU - Magassouba, Aly
AU - Sugiura, Komei
AU - Kawai, Hisashi
N1 - Funding Information:
This work was partially supported by JST CREST, SCOPE and
Funding Information:
This work was partially supported by JST CREST, SCOPE and NEDO.
Publisher Copyright:
© 2020, Springer Nature Switzerland AG.
PY - 2020
Y1 - 2020
N2 - This is an extension from a selected paper from JSAI2019. In this paper, we study data augmentation for visually-grounded language understanding in the context of picking task. A typical picking task consists of predicting a target object specified by an ambiguous instruction,e.g., “Pick up the yellow toy near the bottle”. We specifically show that existing methods for understanding such an instruction can be improved by data augmentation. More explicitly, MCTM [1] and MTCM-GAN [2] show better results with data augmentation when specifically considering latent space features instead of raw features. Additionally our results show that latent-space data augmentation can improve better a network accuracy than regularization methods.
AB - This is an extension from a selected paper from JSAI2019. In this paper, we study data augmentation for visually-grounded language understanding in the context of picking task. A typical picking task consists of predicting a target object specified by an ambiguous instruction,e.g., “Pick up the yellow toy near the bottle”. We specifically show that existing methods for understanding such an instruction can be improved by data augmentation. More explicitly, MCTM [1] and MTCM-GAN [2] show better results with data augmentation when specifically considering latent space features instead of raw features. Additionally our results show that latent-space data augmentation can improve better a network accuracy than regularization methods.
KW - Domestic service robots
KW - Multimodal language understanding
UR - http://www.scopus.com/inward/record.url?scp=85080960873&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85080960873&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-39878-1_17
DO - 10.1007/978-3-030-39878-1_17
M3 - Conference contribution
AN - SCOPUS:85080960873
SN - 9783030398774
T3 - Advances in Intelligent Systems and Computing
SP - 179
EP - 187
BT - Advances in Artificial Intelligence - Selected Papers from the Annual Conference of Japanese Society of Artificial Intelligence JSAI 2019
A2 - Ohsawa, Yukio
A2 - Yada, Katsutoshi
A2 - Ito, Takayuki
A2 - Takama, Yasufumi
A2 - Sato-Shimokawara, Eri
A2 - Abe, Akinori
A2 - Mori, Junichiro
A2 - Matsumura, Naohiro
PB - Springer
T2 - 33rd Annual Conference of the Japanese Society for Artificial Intelligence, JSAI 2019
Y2 - 4 June 2019 through 7 June 2019
ER -