Pragmatic factors in image description: the case of negations
This work addresses the challenge of generating subjective language like negations in image descriptions for AI systems, but it is incremental as it focuses on analysis and pilot annotation without implementing a full system.
The paper analyzed negation usage in the Flickr30K image description corpus, categorizing negation types and proposing requirements for image description systems to generate negation sentences, with a manual annotation pilot achieving an agreement score of K=0.67.
We provide a qualitative analysis of the descriptions containing negations (no, not, n't, nobody, etc) in the Flickr30K corpus, and a categorization of negation uses. Based on this analysis, we provide a set of requirements that an image description system should have in order to generate negation sentences. As a pilot experiment, we used our categorization to manually annotate sentences containing negations in the Flickr30K corpus, with an agreement score of K=0.67. With this paper, we hope to open up a broader discussion of subjective language in image descriptions.