Fine-Grained Semantically Aligned Vision-Language Pre-Training
This addresses the limitation of existing methods that rely on global alignment, potentially improving vision-language models for applications like image captioning and visual question answering, though it is incremental in focusing on fine-grained semantics.
The paper tackles the problem of fine-grained semantic alignment between visual regions and textual phrases in vision-language pre-training, introducing LOUPE which uses game-theoretic interactions to achieve state-of-the-art performance on various tasks and competitive results on object detection and visual grounding without human annotations.
Large-scale vision-language pre-training has shown impressive advances in a wide range of downstream tasks. Existing methods mainly model the cross-modal alignment by the similarity of the global representations of images and texts, or advanced cross-modal attention upon image and text features. However, they fail to explicitly learn the fine-grained semantic alignment between visual regions and textual phrases, as only global image-text alignment information is available. In this paper, we introduce LOUPE, a fine-grained semantically aLigned visiOn-langUage PrE-training framework, which learns fine-grained semantic alignment from the novel perspective of game-theoretic interactions. To efficiently compute the game-theoretic interactions, we further propose an uncertainty-aware neural Shapley interaction learning module. Experiments show that LOUPE achieves state-of-the-art performance on a variety of vision-language tasks. Furthermore, without any object-level human annotations and fine-tuning, LOUPE achieves competitive performance on object detection and visual grounding. More importantly, LOUPE opens a new promising direction of learning fine-grained semantics from large-scale raw image-text pairs. The repository of this work is at https://github.com/YYJMJC/LOUPE.