Xiaoqi Jackie
Zhao

My long-term research vision is to build a general-purpose visual intelligence engine under the AI-for-X paradigm. This vision is structured around three interconnected layers: Pattern, Unification, AI4X Applications. At the pattern level, I investigate self-driven learning mechanisms such as self-supervised, open-vocabulary, prompt-based, and in-context learning to establish scalable and transferable learning principles that reduce supervision dependency and enhance generalization. At the unification level, I study how diverse visual tasks, concepts, and modalities can be integrated within shared architectures and representations, enabling unified multi-task modeling, context-dependent concept understanding, and image-video multi-modal perception under a coherent semantic framework. Built upon these foundations, the AI4X application layer extends the engine into real-world domains such as healthcare and industry, validating its robustness in complex, domain-specific environments while continuously expanding the scope of X. Recently, I have developed a strong interest in AI4ResearchTools, and I warmly welcome collaboration and mentorship.