Measuring Diversity in Synthetic Datasets

Yuchang Zhu, Huizhe Zhang, Bingzhe Wu, Jintang Li, Zibin Zheng, Peilin Zhao, Liang Chen, Yatao Bian

Project

• Proposes a classification-based method to evaluate the diversity of datasets generated by LLMs.

One Fits All: Learning Fair Graph Neural Networks for Various Sensitive Attributes

Yuchang Zhu, Jintang Li, Yatao Bian, Zibin Zheng, Liang Chen

Project

• Presents a causal-inspired group fairness framework for Graph Neural Networks (GNNs) to address fairness issues under multiple sensitive attributes.

Fair Graph Representation Learning via Sensitive Attribute Disentanglement

Yuchang Zhu, Jintang Li, Zibin Zheng, Liang Chen

Project

• Introduces a disentanglement-based approach to learn fair graph representations, enhancing GNN fairness without sacrificing model utility.

The devil is in the data: Learning fair graph neural networks via partial knowledge distillation

Yuchang Zhu, Jintang Li, Liang Chen, Zibin Zheng

Project

• Develops a group fairness method for GNNs using knowledge distillation, effective even when sensitive attributes are entirely unknown.

Fairagg: Toward fair graph neural networks via fair aggregation

Yuchang Zhu, Jintang Li, Liang Chen, Zibin Zheng

Project

• Proposes a novel aggregation scheme for GNNs specifically designed to enhance group fairness during message passing.

SaGIF: Improving Individual Fairness in Graph Neural Networks via Similarity Encoding

Yuchang Zhu, Jintang Li, Huizhe Zhang, Liang Chen, Zibin Zheng

Project

• Introduces an individual fairness method for GNNs that encodes node similarity to ensure that similar individuals receive similar outcomes.

What Matters in LLM-generated Data: Diversity and Its Effect on Model Fine-Tuning

Yuchang Zhu*, Huazhen Zhong*, Qunshu Lin*, Haotong Wei, Xiaolong Sun, Zixuan Yu, Minghao Liu, Zibin Zheng, Liang Chen

Project

• Investigates the role of diversity in data generated by LLMs and its subsequent effect on the performance of fine-tuned models.