Search Results

Documents authored by Mai, Gengchen

Document
DOI: 10.4230/LIPIcs.GIScience.2025.8
BERT4Traj: Transformer-Based Trajectory Reconstruction for Sparse Mobility Data

Authors: Hao Yang, Angela Yao, Christopher C. Whalen, and Gengchen Mai

Published in: LIPIcs, Volume 346, 13th International Conference on Geographic Information Science (GIScience 2025)

Abstract
Understanding human mobility is essential for applications in public health, transportation, and urban planning. However, mobility data often suffers from sparsity due to limitations in data collection methods, such as infrequent GPS sampling or call detail record (CDR) data that only capture locations during communication events. To address this challenge, we propose BERT4Traj, a transformer-based model that reconstructs complete mobility trajectories by predicting hidden visits in sparse movement sequences. Inspired by BERT’s masked language modeling objective and self-attention mechanisms, BERT4Traj leverages spatial embeddings, temporal embeddings, and contextual background features such as demographics and anchor points. We evaluate BERT4Traj on real-world CDR and GPS datasets collected in Kampala, Uganda, demonstrating that our approach significantly outperforms traditional models such as Markov Chains, KNN, RNNs, and LSTMs. Our results show that BERT4Traj effectively reconstructs detailed and continuous mobility trajectories, enhancing insights into human movement patterns.
Cite as

Hao Yang, Angela Yao, Christopher C. Whalen, and Gengchen Mai. BERT4Traj: Transformer-Based Trajectory Reconstruction for Sparse Mobility Data. In 13th International Conference on Geographic Information Science (GIScience 2025). Leibniz International Proceedings in Informatics (LIPIcs), Volume 346, pp. 8:1-8:9, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2025)

Copy BibTex To Clipboard

@InProceedings{yang_et_al:LIPIcs.GIScience.2025.8,
  author =	{Yang, Hao and Yao, Angela and Whalen, Christopher C. and Mai, Gengchen},
  title =	{{BERT4Traj: Transformer-Based Trajectory Reconstruction for Sparse Mobility Data}},
  booktitle =	{13th International Conference on Geographic Information Science (GIScience 2025)},
  pages =	{8:1--8:9},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-378-2},
  ISSN =	{1868-8969},
  year =	{2025},
  volume =	{346},
  editor =	{Sila-Nowicka, Katarzyna and Moore, Antoni and O'Sullivan, David and Adams, Benjamin and Gahegan, Mark},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.GIScience.2025.8},
  URN =		{urn:nbn:de:0030-drops-238373},
  doi =		{10.4230/LIPIcs.GIScience.2025.8},
  annote =	{Keywords: Human Mobility, Trajectory Reconstruction, Deep Learning, CDR, GPS}
}
Document
DOI: 10.4230/LIPIcs.COSIT.2024.9
Probing the Information Theoretical Roots of Spatial Dependence Measures

Authors: Zhangyu Wang, Krzysztof Janowicz, Gengchen Mai, and Ivan Majic

Published in: LIPIcs, Volume 315, 16th International Conference on Spatial Information Theory (COSIT 2024)

Abstract
Intuitively, there is a relation between measures of spatial dependence and information theoretical measures of entropy. For instance, we can provide an intuition of why spatial data is special by stating that, on average, spatial data samples contain less than expected information. Similarly, spatial data, e.g., remotely sensed imagery, that is easy to compress is also likely to show significant spatial autocorrelation. Formulating our (highly specific) core concepts of spatial information theory in the widely used language of information theory opens new perspectives on their differences and similarities and also fosters cross-disciplinary collaboration, e.g., with the broader AI/ML communities. Interestingly, however, this intuitive relation is challenging to formalize and generalize, leading prior work to rely mostly on experimental results, e.g., for describing landscape patterns. In this work, we will explore the information theoretical roots of spatial autocorrelation, more specifically Moran’s I, through the lens of self-information (also known as surprisal) and provide both formal proofs and experiments.
Cite as

Zhangyu Wang, Krzysztof Janowicz, Gengchen Mai, and Ivan Majic. Probing the Information Theoretical Roots of Spatial Dependence Measures. In 16th International Conference on Spatial Information Theory (COSIT 2024). Leibniz International Proceedings in Informatics (LIPIcs), Volume 315, pp. 9:1-9:18, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2024)

Copy BibTex To Clipboard

@InProceedings{wang_et_al:LIPIcs.COSIT.2024.9,
  author =	{Wang, Zhangyu and Janowicz, Krzysztof and Mai, Gengchen and Majic, Ivan},
  title =	{{Probing the Information Theoretical Roots of Spatial Dependence Measures}},
  booktitle =	{16th International Conference on Spatial Information Theory (COSIT 2024)},
  pages =	{9:1--9:18},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-330-0},
  ISSN =	{1868-8969},
  year =	{2024},
  volume =	{315},
  editor =	{Adams, Benjamin and Griffin, Amy L. and Scheider, Simon and McKenzie, Grant},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.COSIT.2024.9},
  URN =		{urn:nbn:de:0030-drops-208247},
  doi =		{10.4230/LIPIcs.COSIT.2024.9},
  annote =	{Keywords: Spatial Autocorrelation, Moran’s I, Information Theory, Surprisal, Self-Information}
}
Document
DOI: 10.4230/LIPIcs.GISCIENCE.2018.17
xNet+SC: Classifying Places Based on Images by Incorporating Spatial Contexts

Authors: Bo Yan, Krzysztof Janowicz, Gengchen Mai, and Rui Zhu

Published in: LIPIcs, Volume 114, 10th International Conference on Geographic Information Science (GIScience 2018)

Abstract
With recent advancements in deep convolutional neural networks, researchers in geographic information science gained access to powerful models to address challenging problems such as extracting objects from satellite imagery. However, as the underlying techniques are essentially borrowed from other research fields, e.g., computer vision or machine translation, they are often not spatially explicit. In this paper, we demonstrate how utilizing the rich information embedded in spatial contexts (SC) can substantially improve the classification of place types from images of their facades and interiors. By experimenting with different types of spatial contexts, namely spatial relatedness, spatial co-location, and spatial sequence pattern, we improve the accuracy of state-of-the-art models such as ResNet - which are known to outperform humans on the ImageNet dataset - by over 40%. Our study raises awareness for leveraging spatial contexts and domain knowledge in general in advancing deep learning models, thereby also demonstrating that theory-driven and data-driven approaches are mutually beneficial.
Cite as

Bo Yan, Krzysztof Janowicz, Gengchen Mai, and Rui Zhu. xNet+SC: Classifying Places Based on Images by Incorporating Spatial Contexts. In 10th International Conference on Geographic Information Science (GIScience 2018). Leibniz International Proceedings in Informatics (LIPIcs), Volume 114, pp. 17:1-17:15, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2018)

Copy BibTex To Clipboard

@InProceedings{yan_et_al:LIPIcs.GISCIENCE.2018.17,
  author =	{Yan, Bo and Janowicz, Krzysztof and Mai, Gengchen and Zhu, Rui},
  title =	{{xNet+SC: Classifying Places Based on Images by Incorporating Spatial Contexts}},
  booktitle =	{10th International Conference on Geographic Information Science (GIScience 2018)},
  pages =	{17:1--17:15},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-083-5},
  ISSN =	{1868-8969},
  year =	{2018},
  volume =	{114},
  editor =	{Winter, Stephan and Griffin, Amy and Sester, Monika},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.GISCIENCE.2018.17},
  URN =		{urn:nbn:de:0030-drops-93450},
  doi =		{10.4230/LIPIcs.GISCIENCE.2018.17},
  annote =	{Keywords: Spatial context, Image classification, Place types, Convolutional neural network, Recurrent neural network}
}
Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail
© 2023-2025 Schloss Dagstuhl – LZI GmbH About DROPS Imprint Privacy Contact