Influence Maximization in Temporal Social Networks with a Cold-start Problem: A Supervised Approach

Laixin Xie, Ying Zhang, Xiyuan Wang, Shiyi Liu, Shenghan Gao, Xingxing Xing, Wei Wan, Haipeng Zhang,Quan Li

Arxiv | Github

• Researched on the cold-start problem in Influence Maximization (IM), cooperating with NetEase.
• Proposed a efficient labeling for the IM seeds, a solution for the cold-start issue and a tensorized TGN for acceleration.
• Employed an online A/B testing in expanding network scale for evaluation.

Deciphering Explicit and Implicit Features for Reliable, Interpretable, and Actionable User Churn Prediction in Online Video Games

Xiyuan Wang* , Laixin Xie* , He Wang , Xingxing Xing , Wei Wan , Ziming Wu , Xiaojuan Ma , and Quan Li (*: equal contribution)

Video | Github

• Researched on the practical player churn prediction, cooperating with NetEase.
• Developed a visual analytic system, Don’tgo to conduct a what-if analysis on the root causes of player churn and to propose actionable interventions to mitigate it.

Towards Better Modeling With Missing Data: A Contrastive Learning-Based Visual Analytics Perspective

Laixin Xie, Yang Ouyang, Longfei Chen, Ziming Wu, Quan Li

Video | Arxiv | Github

• Researched how to leverage contrastive learning to address missing data, cooperating with Tencent.
• Proposed a contrastive learning-based framework, allowing label prediction with the presence of data missing and without any imputation.
• Further developedd a visual analytics system, CIVis to understand the training process of contrastive learning and iterative refinement.

RoleSeer: Understanding Informal Social Role Changes in MMORPGs via Visual Analytics

Laixin Xie, Ziming Wu, Peng Xu, Wei Li, Xiaojuan Ma, Quan Li

Video ｜ Arxiv

• Researched on the informal social role in MMORPGs, cooperating with Tencent and Netease.
• Developed an adapted dynamic network embedding method to identify the potential informal roles from the perspective of behavioral interaction analysis
• Proposed a visual analytics system named RoleSeer to investigate the informal roles from the perspectives of behavioral interactions and depict their dynamic interconversions and transitions.