Fengrui Hao

Style-Based Bias Backdoor Attacks on Medical LLMs Under In-Context Learning

Jiang Y, Shijie Xiao, Jin Lin, Fengrui Hao et autres

cn (code pays fourni par la source)

Fairness-aware differentially private model training without sensitive attributes for face recognition

Fengrui Hao, Yuzhao Chen, Tianlong Gu, Fan Zhang et autres

cn (code pays fourni par la source)

Stealthy backdoor attack method targeting group fairness in self-supervised learning

Fengrui Hao, Tianlong Gu, Jionghui Jiang, Liang Chang et autres

cn (code pays fourni par la source)

FairGSE: Fairness-Aware Graph Neural Network Without High False Positive Rates

Zhenqiang Ye, Jinjie Lu, Tianlong Gu, Fengrui Hao et autres

Graph neural networks (GNNs) have emerged as the mainstream paradigm for graph representation learning due to their effective message aggregation. However, this advantage also amplifies biases inherent in graph topology, raising fairness concerns. Existing fairness-aware GNNs provide satisfactory performance on fairness metrics …

cn (code pays fourni par la source)

DMPA: Durable Model Poisoning Attack Against Fairness and Robustness in Efficient Federated Learning Systems

Jionghui Jiang, Fengrui Hao, Tianlong Gu, Ke Wang et autres

Federated Learning (FL) systems are increasingly deployed across multiple clients to efficiently train a shared model over local data, thereby effectively addressing data silos and reducing communication. However, FL systems are known to be susceptible to model poisoning attacks by malicious clients, …

cn (code pays fourni par la source)

Fairness-aware Differentially Private Model Training without Sensitive Attributes for Face Recognition

Fengrui Hao, Yuzhao Chen, Tianlong Gu, Fan Zhang et autres

cn (code pays fourni par la source)

FairDP-GNN: Graph Neural Network with Group Fairness and Differential Privacy

Fengrui Hao, Shiyi Zhao, Tianlong Gu, Xuemin Wang et autres

cn, jp (code pays fourni par la source)

C2PO: Diagnosing and Disentangling Bias Shortcuts in LLMs

Xuan Feng, Bo An, Tianlong Gu, Liang Chang et autres

Bias in Large Language Models (LLMs) poses significant risks to trustworthiness, manifesting primarily as stereotypical biases (e.g., gender or racial stereotypes) and structural biases (e.g., lexical overlap or position preferences).However, prior paradigms typically address these in isolation, often mitigating one at the …

sg, cn (code pays fourni par la source)

C2PO: Diagnosing and Disentangling Bias Shortcuts in LLMs

Xuan Feng, Bo An, Tianlong Gu, Liang Chang et autres

Bias in Large Language Models (LLMs) poses significant risks to trustworthiness, manifesting primarily as stereotypical biases (e.g., gender or racial stereotypes) and structural biases (e.g., lexical overlap or position preferences). However, prior paradigms typically address these in isolation, often mitigating one at …

C2PO: Diagnosing and Disentangling Bias Shortcuts in LLMs

Xuan Feng, Bo An, Tianlong Gu, Liang Chang et autres

Bias in Large Language Models (LLMs) poses significant risks to trustworthiness, manifesting primarily as stereotypical biases (e.g., gender or racial stereotypes) and structural biases (e.g., lexical overlap or position preferences). However, prior paradigms typically address these in isolation, often mitigating one at …

sg, cn (code pays fourni par la source)

Exacerbating Differences in Polarity: Bias Adversarial Attack on Generative Large Language Models

Tianlong Gu, Fengrui Hao, Liang Chang

Generative large language models (LLMs) have demonstrated outstanding performance across a wide range of language-related tasks. However, these models may inherit or even amplify biases in their training data, potentially leading to adverse impacts on specific individuals or groups. To this end, …

cn (code pays fourni par la source)

On the correlations between geometric metrics and fairness in pruning CNN

Xiaoli Liu, Fengrui Hao

cn (code pays fourni par la source)