Zhao-Rong Lai

$\ell_1$ Latent Distance based Continuous-time Graph Representation

A Unified Total Variation Framework for Membrane Potential Perturbation Dynamic

Functional Shrinkage for Input Optimization

Non-differentiable Regularization for Heavy-tailed Differentially Private Stochastic Convex Optimization

Adversarial Attack and Defense for Denoising Diffusion Sampling

A Proximal Stochastic Gradient Method for Doubly-regularized Spectral Risk Minimization

Privacy Amplification by Iteration with Projected Alternating Direction Method

Spectral Decomposed Variational Inference: A Principled Framework for Posterior Covariance Modeling

Understanding Hardness of Vision-Language Compositionality from A Token-level Causal Lens

Quadratic Coreset Selection: Certifying and Reconciling Sequence and Token Mining for Efficient Instruction Tuning

Language Models as Implicit Tree Search

Out-of-distribution Generalization for Total Variation based Invariant Risk Minimization