Dong-Jun Han

Assistant Professor@Yonsei University·韩国·OpenReview

研究方向

Model Compression · Efficient Training Algorithms · Active Learning · Continual Learning · Out-of-Distribution · Federated Learning · Distributed Machine Learning

Dong-Jun Han

Federated Sketching LoRA: A Flexible Framework for Heterogeneous Collaborative Fine-Tuning of LLMs

TAP: Two-Stage Adaptive Personalization of Multi-task and Multi-Modal Foundation Models in Federated Learning

Identifying Robust Neural Pathways: Few-Shot Adversarial Mask Tuning for Vision-Language Models

Gradient Correction in Federated Learning with Adaptive Optimization

Reshape-then-Factorize: Communication-Efficient FL via Model-Agnostic Projection Optimization

Collaborative Device-Cloud LLM Inference through Reinforcement Learning

Elevating the Tradeoff between Privacy and Utility in Zeroth-Order Vertical Federated Learning

Decentralized Sporadic Federated Learning: A Unified Algorithmic Framework with Convergence Guarantees

Reshape-then-Factorize: Communication-Efficient FL via Model-Agnostic Projection Optimization

Adaptive Energy Alignment for Accelerating Test-Time Adaptation

Unlocking the Potential of Model Calibration in Federated Learning

Asynchronous Federated Reinforcement Learning with Policy Gradient Updates: Algorithm Design and Convergence Analysis

PRISM: Privacy-Preserving Improved Stochastic Masking for Federated Generative Models

SePPO: Semi-Policy Preference Optimization for Diffusion Alignment

Communication-Efficient Federated Learning via Model-Agnostic Projection Adaptation

Communication Efficient Federated Learning via Model-Agnostic Projection Adaptation

Federated Learning with Dynamic Client Arrival and Departure: Convergence and Rapid Adaptation via Initial Model Construction