12 papers
Safety Anchor: Defending Harmful Fine-tuning via Geometric Bottlenecks
Guoxin Lu, Letian Sha, Qing Wang +4
The safety alignment of Large Language Models (LLMs) remains vulnerable to Harmful Fine-tuning (HFT). While existing defenses impose constraints on parameters, gradients, or intern…
Conceal, Reconstruct, Jailbreak: Exploiting the Reconstruction-Concealment Tradeoff in MLLMs
Md Farhamdur Reza, Richeng Jin, Tianfu Wu +1
Intent-obfuscation-based jailbreak attacks on multimodal large language models (MLLMs) transform a harmful query into a concealed multimodal input to bypass safety mechanisms. We s…
Mobility-Assisted Decentralized Federated Learning: Convergence Analysis and A Data-Driven Approach
Reza Jahani, Md Farhamdur Reza, Richeng Jin +1
Decentralized Federated Learning (DFL) has emerged as a privacy-preserving machine learning paradigm that enables collaborative training among users without relying on a central se…
Gradient Compression May Hurt Generalization: A Remedy by Synthetic Data Guided Sharpness Aware Minimization
Yujie Gu, Richeng Jin, Zhaoyang Zhang +1
It is commonly believed that gradient compression in federated learning (FL) enjoys significant improvement in communication efficiency with negligible performance degradation. In…
Computation- and Communication-Efficient Online FL for Resource-Constrained Aerial Vehicles
Ferdous Pervej, Richeng Jin, Md Moin Uddin Chowdhury +3
Privacy-preserving distributed machine learning (ML) and aerial connected vehicle (ACV)-assisted edge computing have drawn significant attention lately. Since the onboard sensors o…
Byzantine Outside, Curious Inside: Reconstructing Data Through Malicious Updates
Kai Yue, Richeng Jin, Chau-Wai Wong +1
Federated learning (FL) enables decentralized machine learning without sharing raw data, allowing multiple clients to collaboratively learn a global model. However, studies reveal…