Xiujian Liang梁秀健

我是 复旦大学（Fudan University）计算机科学与技术直博生，在 上海创智学院（Shanghai Innovation Institute, SII） 联合培养，现由 卢策吾教授 指导。我的研究聚焦 可靠机器人感知、几何约束世界模型与机器人操作学习，连接真实场景中的 Sim-to-Real 感知、空间状态表征和动作条件动态预测，探索如何让机器人理解物理环境并可靠地预测动作后果。目前，我正在进一步研究显式三维几何与可学习动态潜变量相结合的 Geometry-Grounded World-Action Models，面向物理交互预测、机器人决策与操作泛化。I am a Computer Science Ph.D. student at Fudan University, jointly trained at the Shanghai Innovation Institute (SII), advised by Prof. Cewu Lu. My research centers on Reliable Robotic Perception, Geometry-Grounded World Models, and Robot Learning for Manipulation, connecting sim-to-real sensing, spatial state representations, and action-conditioned dynamics to help robots anticipate the consequences of physical interaction. I am currently exploring Geometry-Grounded World-Action Models that combine explicit 3D geometry with learned latent dynamics for reliable action evaluation, planning, and generalization.