LT3 at SemEval-2020 Task 8 : multi-modal multi-task learning for memotion analysis

Publication type
C1
Publication status
Published
Authors
Singh, P., Bauwelinck, N.B., & Lefever, E.
Series
Proceedings of the 14th International Workshop on Semantic Evaluation (SemEval 2020)
Pagination
1155-1162
Publisher
Association for Computational Linguistics (ACL) (Barcelona, Spain)
Conference
the Fourteenth Workshop on Semantic Evaluation (SemEval 2020) (Barcelona, Spain (online))
Download
(.pdf)
View in Biblio
(externe link)

Abstract

Internet memes have become a very popular mode of expression on social media networks today.
Their multi-modal nature, caused by a mixture of text and image, makes them a very challenging
research object for automatic analysis. In this paper, we describe our contribution to the SemEval2020 Memotion Analysis Task. We propose a Multi-Modal Multi-Task learning system, which
incorporates “memebeddings”, viz. joint text and vision features, to learn and optimize for all
three Memotion subtasks simultaneously. The experimental results show that the proposed system
constantly outperforms the competition’s baseline, and the system setup with continual learning
(where tasks are trained sequentially) obtains the best classification F1-scores.