Towards Zero-shot Cross-lingual Image Retrieval
This work is significant for researchers and practitioners working on multi-modal language and vision problems, particularly those aiming to extend models beyond English to support diverse linguistic communities, by providing a method for zero-shot cross-lingual image retrieval and a new evaluation dataset.
This paper addresses the problem of cross-lingual image retrieval by proposing a zero-shot approach that trains on monolingual data but can be applied to multiple languages during inference. They introduce a new objective function to improve text embedding clusters and create a new 1K multi-lingual MSCOCO2014 caption test dataset (XTD10) in 7 languages for evaluation.
There has been a recent spike in interest in multi-modal Language and Vision problems. On the language side, most of these models primarily focus on English since most multi-modal datasets are monolingual. We try to bridge this gap with a zero-shot approach for learning multi-modal representations using cross-lingual pre-training on the text side. We present a simple yet practical approach for building a cross-lingual image retrieval model which trains on a monolingual training dataset but can be used in a zero-shot cross-lingual fashion during inference. We also introduce a new objective function which tightens the text embedding clusters by pushing dissimilar texts from each other. Finally, we introduce a new 1K multi-lingual MSCOCO2014 caption test dataset (XTD10) in 7 languages that we collected using a crowdsourcing platform. We use this as the test set for evaluating zero-shot model performance across languages. XTD10 dataset is made publicly available here: https://github.com/adobe-research/Cross-lingual-Test-Dataset-XTD10