Harmonizing Visual Representations for Unified Multimodal Understanding and Generation Published: September 17, 2026Share on Twitter Facebook Google+ LinkedIn Previous Next