Bin Lu (卢彬)

Logo

PhD Student
Department of Electronic Engineering
Shanghai Jiao Tong University

Office: 441 SEIEE No.1 Building
Email: robinlu1209@sjtu.edu.cn

Google Scholar
DBLP
Semantic Scholar
Github

View My GitHub Profile

Bin Lu’s homepage

Biography

I’m currently a final year PhD student in IIOT Research Center at Shanghai Jiao Tong University, advised by Professor Xiaoying Gan and Professor Xinbing Wang. I’m also a member of Wen-Tsun Wu AI honorary doctoral class. I received bachelor from Shanghai Jiao Tong University with honors in July 2020 and was recommended for the master degree from Sepetember 2020. Additionally, I transfer to be a PhD student (硕博连读) in March, 2022.

My current research interests focus on revealing heterogeneous structure of ubiquitous entities and discovering underlying knowledge with machine learning techniques.

research_interest

Publications

[ICML 2024] OxyGenerator: Reconstructing Global Ocean Deoxygenation Over a Century with Deep Learning [ICLR Workshop 2024][EGU]
Bin Lu, Ze Zhao, Luyu Han, Xiaoying Gan, Yuntao Zhou, Lei Zhou, Luoyi Fu, Xinbing Wang, Chenghu Zhou, Jing Zhang
International Conference on Machine Learning, 2024. (CCF-A)

[TKDE 2024] Graph Out-of-Distribution Generalization with Controllable Data Augmentation
Bin Lu, Ze Zhao, Xiaoying Gan, Shiyu Liang, Luoyi Fu, Xinbing Wang, Chenghu Zhou
IEEE Transactions on Knowledge and Data Engineering, 2024. (CCF-A)

[ICLR 2024] Temporal Generalization Estimation in Evolving Graphs [paper][NeurIPS Workshop 2023]
Bin Lu, Tingyan Ma, Xiaoying Gan, Xinbing Wang, Yunqiang Zhu, Chenghu Zhou, Shiyu Liang
The Twelfth International Conference on Learning Representations, 2024. (Tsinghua-A, the short version is presented in NeurIPS 2023 Workshop on Self-Supervised Learning - Theory and Practice.)

[WWW 2023] DataExpo: A One-stop Dataset Service for Open Science Research [demo][paper]
Bin Lu, Lynwen Wu, Lina Yang, Chenxing Sun, Wei Liu, Xiaoying Gan, Shiyu Liang, Xinbing Wang, Chenghu Zhou
The Web Conference, 2023. (CCF-A)

[KDD 2022] Geometer: Graph Few-Shot Class-Incremental Learning via Prototype Representation [arxiv][code][paper][slide][poster][media]
Bin Lu, Xiaoying Gan, Lina Yang, Weinan Zhang, Luoyi Fu, Xinbing Wang
ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022. (CCF-A, Oral)

[KDD 2022] Spatio-Temporal Graph Few-Shot Learning with Cross-City Knowledge Transfer [arxiv][code][paper][slide][poster][media]
Bin Lu, Xiaoying Gan, Weinan Zhang, Huaxiu Yao, Luoyi Fu, Xinbing Wang
ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2022. (CCF-A, Oral)

[ACM TIST 2022] Make More Connections: Urban Traffic Flow Forecasting with Spatiotemporal Adaptive Gated Graph Convolution Network [paper]
Bin Lu, Xiaoying Gan, Haiming Jin, Luoyi Fu, Xinbing Wang, Haisong Zhang
ACM Transactions on Intelligent Systems and Technology, 2022. (SCI，SJTU-A，IF=4.654)

[CIKM 2020] Spatiotemporal Adaptive Gated Graph Convolution Network for Urban Traffic Flow Forecasting [paper][code][slide]
Bin Lu, Xiaoying Gan, Haiming Jin, Luoyi Fu, Haisong Zhang
ACM International Conference on Information and Knowledge Management, 2020. (CCF-B, Oral, included in open-source project LibCity)

[ICDM 2023] Graph Open-Set Recognition via Entropy Message Passing
Lina Yang, Bin Lu, Xiaoying Gan
IEEE International Conference on Data Mining, 2023. (CCF-B)

[SIGIR 2024] EditKG: Editing Knowledge Graph for Recommendation
Gu Tang, Xiaoying Gan, Jinghe Wang, Bin Lu, Lyuwen Wu, Luoyi Fu, Chenghu Zhou
International ACM SIGIR Conference on Research and Development in Information Retrieval, 2024. (CCF-A)

Project

DDE DataExpo – Deep-time Digital Earth (DDE) Program. The Deep-time Digital Earth (DDE) Program is the first ‘big science program’ initiated by the International Union of Geological Sciences (IUGS) that will provide new opportunities and directions for the development of Earth Sciences. DDE DataStore is an important component of Deep-time Digital Earth (DDE) Program, aiming to discover and integrate global geological data. At present, DDE DataExpo has obtained more than 1 million websites based on 15,000 keywords and found more than 200,000 geological data.[website][poster][media][news]

dde datastore

Jingwei - Hybrid Graph Learning Reconstructs Global Ocean Changes. In this interdisciplinary study, we present a significant achievement in the modeling of ocean deoxygenation, offering accurate, long-term, and global spatiotemporal reconstructions of dissolved oxygen variations through a novel hybrid deep learning framework. We name it Jingwei, a mythical bird in Chinese mythology that insist on filling the sea with pebbles and twigs, just as we try to fill the sparse ocean oxygen observations. We believe the approach proposed in this study represents a new AI-driven paradigm in ocean system modeling, with the potential to advance the understanding of deoxygenation mechanisms and provide a pioneering framework for climate change analysis with advanced AI techniques. Alongside, we have released an open-source platform, providing data visualization, resource sharing and ongoing updates to more variables. [code] [website]

jingwei framework

Awards

Grants

Professional Activities

Program Committee Member

Conference Reviewer

Journal Reviewer

Presentation

Teaching Assistant

Competitions

Contact

Room 441, SEIEE Building No.1, Dongchuan Road 800, Shanghai, China Email: robinlu1209 at sjtu dot edu dot cn