Negative Sample is Negative in Its Own Way: Tailoring Negative Sentences for Image-Text Retrieval

Zhihao Fan, Zhongyu Wei, Zejun Li, Siyuan Wang, Xuanjing Huang, Jianqing Fan


Abstract
Matching model is essential for Image-Text Retrieval framework. Existing research usually train the model with a triplet loss and explore various strategy to retrieve hard negative sentences in the dataset. We argue that current retrieval-based negative sample construction approach is limited in the scale of the dataset thus fail to identify negative sample of high difficulty for every image. We propose our TAiloring neGative Sentences with Discrimination and Correction (TAGS-DC) to generate synthetic sentences automatically as negative samples. TAGS-DC is composed of masking and refilling to generate synthetic negative sentences with higher difficulty. To keep the difficulty during training, we mutually improve the retrieval and generation through parameter sharing. To further utilize fine-grained semantic of mismatch in the negative sentence, we propose two auxiliary tasks, namely word discrimination and word correction to improve the training. In experiments, we verify the effectiveness of our model on MS-COCO and Flickr30K compared with current state-of-the-art models and demonstrates its robustness and faithfulness in the further analysis.
Anthology ID:
2022.findings-naacl.204
Volume:
Findings of the Association for Computational Linguistics: NAACL 2022
Month:
July
Year:
2022
Address:
Seattle, United States
Editors:
Marine Carpuat, Marie-Catherine de Marneffe, Ivan Vladimir Meza Ruiz
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2667–2678
Language:
URL:
https://aclanthology.org/2022.findings-naacl.204
DOI:
10.18653/v1/2022.findings-naacl.204
Bibkey:
Cite (ACL):
Zhihao Fan, Zhongyu Wei, Zejun Li, Siyuan Wang, Xuanjing Huang, and Jianqing Fan. 2022. Negative Sample is Negative in Its Own Way: Tailoring Negative Sentences for Image-Text Retrieval. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 2667–2678, Seattle, United States. Association for Computational Linguistics.
Cite (Informal):
Negative Sample is Negative in Its Own Way: Tailoring Negative Sentences for Image-Text Retrieval (Fan et al., Findings 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-2/2022.findings-naacl.204.pdf
Video:
 https://preview.aclanthology.org/nschneid-patch-2/2022.findings-naacl.204.mp4
Code
 libertfan/tags
Data
MS COCO