Jiaru Zhang

FedMomentum: Preserving LoRA Training Momentum in Federated Fine-Tuning

Test Time Training for Supervised Causal Learning

Leveraging Model Guidance to Extract Training Data from Personalized Diffusion Models

Single-Step Consistent Diffusion Samplers

Revealing the Unseen: Guiding Personalized Diffusion Models to Expose Training Data

Exploring Diffusion Models' Corruption Stage in Few-Shot Fine-tuning and Mitigating with Bayesian Neural Networks