Transferable Multimodal Attack on Vision-Language Pre-training Models | IEEE Conference Publication | IEEE Xplore