Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Dense and Aligned Captions (DAC) Promote Compositional Reasoning in VL Models

About

Vision and Language (VL) models offer an effective method for aligning representation spaces of images and text, leading to numerous applications such as cross-modal retrieval, visual question answering, captioning, and more. However, the aligned image-text spaces learned by all the popular VL models are still suffering from the so-called `object bias' - their representations behave as `bags of nouns', mostly ignoring or downsizing the attributes, relations, and states of objects described/appearing in texts/images. Although some great attempts at fixing these `compositional reasoning' issues were proposed in the recent literature, the problem is still far from being solved. In this paper, we uncover two factors limiting the VL models' compositional reasoning performance. These two factors are properties of the paired VL dataset used for finetuning and pre-training the VL model: (i) the caption quality, or in other words `image-alignment', of the texts; and (ii) the `density' of the captions in the sense of mentioning all the details appearing on the image. We propose a fine-tuning approach for automatically treating these factors leveraging a standard VL dataset (CC3M). Applied to CLIP, we demonstrate its significant compositional reasoning performance increase of up to $\sim27\%$ over the base model, up to $\sim20\%$ over the strongest baseline, and by $6.7\%$ on average.

Sivan Doveh, Assaf Arbelle, Sivan Harary, Roei Herzig, Donghyun Kim, Paola Cascante-bonilla, Amit Alfassy, Rameswar Panda, Raja Giryes, Rogerio Feris, Shimon Ullman, Leonid Karlinsky• 2023

Related benchmarks

TaskDatasetResultRank
Image ClassificationImageNet-1K
Top-1 Acc51.02
524
Text RetrievalFlickr30K--
75
Compositional ReasoningSugarCrepe
Overall Accuracy86.41
43
Compositional ReasoningVL-Checklist
Attribute Score77.27
37
Text RetrievalMS-COCO--
37
Image RetrievalMS-COCO
R@563.51
36
Image RetrievalFlickr30K
Recall@587.84
21
Compositional ReasoningARO
Relation Score81.3
17
Vision-Language ProbingVL-CheckList (test)
Object: Avg77.3
17
Showing 9 of 9 rows

Other info

Follow for update