科学研究
结构表征学习实验室
PI：Yue SONG
研究方向：面向科学智能的结构化表征学习：双向赋能AI与科学发现
课题组简介
我们课题组致力于在科学与人工智能的交叉前沿开展结构化表征学习研究。我们的核心研究范式是：从物理学、神经科学等自然科学领域所蕴含的结构化规律中（如矩阵几何、动力系统、振荡同步机制等），提炼出具有普适性的信息先验与约束条件，并将其转化为可计算、可嵌入的归纳偏置。以此为基础，我们构建兼具物理一致性、语义解耦性与变换等变性的新一代人工智能模型（Science4AI），显著增强其可解释性、泛化能力与样本效率。最终，我们将这些模型系统应用于应对复杂科学发现中的关键挑战（AI4Science），例如多尺度动态系统的建模与预测、高维逆问题的求解、以及非欧结构数据（如流体场、生物运动轨迹、材料微结构）的生成与推理，推动科学智能新范式的建立。
代表性论文
• Yue Song, T. Anderson Keller, Nicu Sebe, and Max Welling. Structured Representation Learning: From Homomorphisms and Disentanglement to Equivariance and Topography. Springer Nature. 2025.
• Yue Song, T. Anderson Keller, Nicu Sebe, and Max Welling. “Flow Factorized Representation Learning”. NeurIPS. 2023.
• Yue Song, Nicu Sebe, and Wei Wang. “Fast Differentiable Matrix Square Root”. ICLR. 2022.
• Yue Song, T. Anderson Keller, Sevan Brodjian, Takeru Miyato, Yisong Yue, Pietro Perona, and Max Welling. “Kuramoto Orientation Diffusion Models”. NeurIPS. 2025.
• Yue Song, T. Anderson Keller, Nicu Sebe, and Max Welling. “Latent Traversals in Generative Models as Potential Flows”. ICML. 2023.
• Yue Song, T. Anderson Keller, Yisong Yue, Pietro Perona, and Max Welling. “Langevin Flows for Modeling Neural Latent Dynamics”. Cognitive Computational Neuroscience (CCN). 2025.
• Yue Song, Nicu Sebe, and Wei Wang. “Rankfeat: Rank-1 feature removal for out-of-distribution detection”. NeurIPS. 2022.
• Ziheng Chen, Yue Song*, Yunmei Liu, and Nicu Sebe. “A Lie Group Approach to Riemannian Batch Normalization”. ICLR. 2024. (*denotes corresponding author)
• Guanghao Wei, Yining Huang, Chenru Duan, Yue Song†, and Yuanqi Du†. “Navigating Chemical Space with Latent Flows”. NeurIPS. 2024. († denotes co-supervision).
• Raphi Kang, Yue Song, Georgia Gkioxari, and Pietro Perona. “Is CLIP ideal? No. Can we fix it? Yes!” ICCV. 2025.