Kaleido-BERT: Vision-Language Pre-training on Fashion Domain
This work addresses the need for better cross-modality representations in fashion e-commerce, offering incremental but strong gains over existing methods.
The paper tackled the problem of vision-language pre-training in the fashion domain by introducing Kaleido-BERT with a novel kaleido strategy and alignment guided masking, achieving state-of-the-art results with absolute improvements such as 7.13% in image retrieval R@1 and 3.28% in category recognition accuracy.
We present a new vision-language (VL) pre-training model dubbed Kaleido-BERT, which introduces a novel kaleido strategy for fashion cross-modality representations from transformers. In contrast to random masking strategy of recent VL models, we design alignment guided masking to jointly focus more on image-text semantic relations. To this end, we carry out five novel tasks, i.e., rotation, jigsaw, camouflage, grey-to-color, and blank-to-color for self-supervised VL pre-training at patches of different scale. Kaleido-BERT is conceptually simple and easy to extend to the existing BERT framework, it attains new state-of-the-art results by large margins on four downstream tasks, including text retrieval (R@1: 4.03% absolute improvement), image retrieval (R@1: 7.13% abs imv.), category recognition (ACC: 3.28% abs imv.), and fashion captioning (Bleu4: 1.2 abs imv.). We validate the efficiency of Kaleido-BERT on a wide range of e-commerical websites, demonstrating its broader potential in real-world applications.