Tieliang Gong

S$^2$MAM: Semi-supervised Meta Additive Model for Robust Estimation and Variable Selection

Learning Once, Routing Right: Information-Theoretic Gating for Online Continual Mixture-of-Experts

Exactly Tight Information-theoretic Generalization Bounds via Binary Jensen-Shannon Divergence

InfoSAM: Fine-Tuning the Segment Anything Model from An Information-Theoretic Perspective

Towards Generalization Bounds of GCNs for Adversarially Robust Node Classification

S$^2$MAM: Semi-supervised Meta Additive Model for Robust Estimation and Variable Selection

Rectified Diffusion Guidance for Conditional Generation