Proceedings of the 14th ACM Multimedia Systems Conference 2023
DOI: 10.1145/3587819.3592545
|View full text |Cite
|
Sign up to set email alerts
|

TotalDefMeme: A Multi-Attribute Meme dataset on Total Defence in Singapore

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
3
0

Year Published

2023
2023
2024
2024

Publication Types

Select...
6

Relationship

2
4

Authors

Journals

citations
Cited by 6 publications
(3 citation statements)
references
References 12 publications
0
3
0
Order By: Relevance
“…Importantly, MATK integrates model analysis techniques that offer practitioners and researchers valuable insights into the strengths and weaknesses of their models. Our future plans for MATK include the following: (i) expanding dataset and model support to include the TotalDefMeme dataset [14], MET-Meme dataset [28], and the DisMultiHate model [5]; (ii) enabling multi-task training with different datasets, such as using FHM for hateful classification and HarMeme for harmful classification; (iii) enhancing the documentation to provide comprehensive and detailed information.…”
Section: Discussionmentioning
confidence: 99%
“…Importantly, MATK integrates model analysis techniques that offer practitioners and researchers valuable insights into the strengths and weaknesses of their models. Our future plans for MATK include the following: (i) expanding dataset and model support to include the TotalDefMeme dataset [14], MET-Meme dataset [28], and the DisMultiHate model [5]; (ii) enabling multi-task training with different datasets, such as using FHM for hateful classification and HarMeme for harmful classification; (iii) enhancing the documentation to provide comprehensive and detailed information.…”
Section: Discussionmentioning
confidence: 99%
“…In the context of Singaporean meme analysis, the TotalDefMeme dataset [8], another resource cited on the challenge website, illustrates the complexity of this task. The dataset was annotated by three annotators per meme, revealing significant discrepancies in their interpretations, despite all annotators being from the same cultural background.…”
Section: Harmfulness Classificationmentioning
confidence: 99%
“…The identification of these sensitive topics aids in effective intervention and prevention. • Total Defence Memes (TotalDefMeme) [25]: TotalDefMeme features locally relevant topics related to Singapore's Total Defence concept, providing valuable insights into the associated events and activities.…”
Section: Experiments Settingsmentioning
confidence: 99%