Tao Yan颜韬.

Combining computational neuroscience and brain network dynamics, I study network propagation, control, and EEG representation learning. By integrating neural mechanisms with data-driven learning, I explore interpretable and generalizable foundation models for psychiatric diagnostic support.

I model disease onset, progression, and outcomes by integrating patient states, longitudinal clinical data, and diagnostic evidence. Through intervention modeling and counterfactual analysis, I explore psychiatric diagnostic support, treatment selection, and dynamic clinical decision support.

I study knowledge localization, targeted correction, and continual updates in large models, using knowledge editing for safety governance. Focusing on medical knowledge updates and erroneous, outdated, or risk-relevant knowledge, I investigate consistency, capability preservation, privacy, and reliability after editing.