YIJUN LIN

About Me

I am working with Prof. Yao-Yi Chiang in the Knowledge Computing Lab at UMN CS&E.

I was in the University of Southern California (USC) advised by Prof. Yao-Yi Chiang and Prof. José Luis Ambite.

I am interested in developing machine learning methods for spatiotemporal prediction and forecasting.
My research focuses on incorporate prior knowledge (e.g., spatial properties) to learn representations from sparsely labeled data (e.g., air quality sensor data) to solve various fine-spatial-scale prediction and time-series forecast problems.

➪ More information is available in my CV and on our Lab website.

News

🎉 Our paper "LIGHT: Multi-Modal Text Linking on Historical Maps" is accepted by ICDAR2025 as an oral presentation.

💡 I teach UMN CSCI 5523 Introduction to Data Mining in Spring 2025

Awards

➪ See the full list of awards in CV

🎉 Doctoral Dissertation Fellowship, 2024-2025, The University of Minnesota Graduate School

🎉 UMN DSI-ADC Fellowship, 2022-2024

🎉 First-place, Map Feature Extraction Challenge, AI for Critical Mineral Assessment Competition, 2022, Duan, W., Li, Z., Lin, F., Lin, Y., Shrotriya, T., Knoblock, C. A., Chiang, Y.-Y.

Publications

➪ See the full list of publications in Google Scholar and CV

Lin, Y., Olson, R., Wu, J., Chiang, Y.-Y., and Weinman, J. (2025). LIGHT: Multi-Modal Text Linking on Historical Maps. In Proceedings of the 19th International Conference on Document Analysis and Recognition, Wuhan, Hubei, China (Oral Presentation)

Lin, Y. and Chiang, Y.-Y. (2024). Hyper-Local Deformable Transformers for Text Spotting on Historical Maps. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp. 5387-5397, Barcelona, Spain

Lin, Y. and Chiang, Y.-Y. (2023). Modeling Spatially Varying Physical Dynamics for Spatiotemporal Predictive Learning. In Proceedings of the 31st ACM SIGSPATIAL international conference on advances in geographic information systems, pp. 1-11, Hamburg, Germany

Lin, Y. and Chiang, Y.-Y. (2022). A Semi-Supervised Learning Approach for Abnormal Event Prediction on Large Network Operation Time-Series Data. In Proceedings of the 2022 IEEE International Conference on Big Data, pp. 1024-1033, Osaka, Japan

Lin, Y., Chiang, Y.-Y., Franklin, M., Eckel, S. P. and Ambite, J. L. (2020). Building Autocorrelation-Aware Representations for Fine-Scale Spatiotemporal Prediction. In Proceedings of IEEE International Conference on Data Mining (ICDM), pp. 352-361, Sorrento, Italy (9.8% acceptance rate)