MDL-NAS: A Joint Multi-domain Learning Framework for Vision Transformer | IEEE Conference Publication | IEEE Xplore