Wu Qinghua 吴青桦

WuQinghua is an graduate student at the intersection of brain-computer interfaces and neural engineering. My research interests focus on cross-subject transfer learning, calibration-free decoding, and online test-time adaptation for EEG-based systems. As an entry-level researcher, I am currently learning to integrate domain generalization with lightweight online learning to enable plug-and-play BCIs that can adapt safely without any labeled data. My preliminary studies involve mutual information-guided channel selection, robust pseudo-label filtering, and dual-stage adaptation strategies to address the challenges of weak supervision and domain shift. Through this work, I aim to move toward practical zero-calibration BCIs that work reliably out-of-the-box for new users.