Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Transferable Unsupervised Outlier Detection Framework for Human Semantic Trajectories

Zheng Zhang, Hossein Amiri, Dazhou Yu, Yuntong Hu, Liang Zhao, Andreas Zufle

TL;DR

TOD4Traj presents a transferable unsupervised framework for outlier detection in human semantic trajectories by unifying multi-modal data and leveraging two levels of contrastive learning: modality alignment across spatial-temporal and semantic information, and trajectory-level regularity learning that exploits temporal periodicity. The framework introduces a combined objective $\mathcal{L}=\mathcal{L}_{Align}+\mathcal{L}_{Consistency}+\beta\mathcal{L}_{Clustering}$ to produce embeddings used for cross-time and cross-population anomaly scoring, with low-rank structure captured via a centroid-based soft clustering approach. Experiments on six simulated POL datasets and GeoLife show TOD4Traj outperforms baselines in AUC and Top-K metrics, and demonstrates transferability across cities and robustness to parameter variations. The work provides open datasets and code, highlighting practical potential for real-world surveillance, eldercare, and urban planning tasks where unseen outliers across modalities arise.

Abstract

Semantic trajectories, which enrich spatial-temporal data with textual information such as trip purposes or location activities, are key for identifying outlier behaviors critical to healthcare, social security, and urban planning. Traditional outlier detection relies on heuristic rules, which requires domain knowledge and limits its ability to identify unseen outliers. Besides, there lacks a comprehensive approach that can jointly consider multi-modal data across spatial, temporal, and textual dimensions. Addressing the need for a domain-agnostic model, we propose the Transferable Outlier Detection for Human Semantic Trajectories (TOD4Traj) framework.TOD4Traj first introduces a modality feature unification module to align diverse data feature representations, enabling the integration of multi-modal information and enhancing transferability across different datasets. A contrastive learning module is further pro-posed for identifying regular mobility patterns both temporally and across populations, allowing for a joint detection of outliers based on individual consistency and group majority patterns. Our experimental results have shown TOD4Traj's superior performance over existing models, demonstrating its effectiveness and adaptability in detecting human trajectory outliers across various datasets.

Transferable Unsupervised Outlier Detection Framework for Human Semantic Trajectories

TL;DR

TOD4Traj presents a transferable unsupervised framework for outlier detection in human semantic trajectories by unifying multi-modal data and leveraging two levels of contrastive learning: modality alignment across spatial-temporal and semantic information, and trajectory-level regularity learning that exploits temporal periodicity. The framework introduces a combined objective to produce embeddings used for cross-time and cross-population anomaly scoring, with low-rank structure captured via a centroid-based soft clustering approach. Experiments on six simulated POL datasets and GeoLife show TOD4Traj outperforms baselines in AUC and Top-K metrics, and demonstrates transferability across cities and robustness to parameter variations. The work provides open datasets and code, highlighting practical potential for real-world surveillance, eldercare, and urban planning tasks where unseen outliers across modalities arise.

Abstract

Semantic trajectories, which enrich spatial-temporal data with textual information such as trip purposes or location activities, are key for identifying outlier behaviors critical to healthcare, social security, and urban planning. Traditional outlier detection relies on heuristic rules, which requires domain knowledge and limits its ability to identify unseen outliers. Besides, there lacks a comprehensive approach that can jointly consider multi-modal data across spatial, temporal, and textual dimensions. Addressing the need for a domain-agnostic model, we propose the Transferable Outlier Detection for Human Semantic Trajectories (TOD4Traj) framework.TOD4Traj first introduces a modality feature unification module to align diverse data feature representations, enabling the integration of multi-modal information and enhancing transferability across different datasets. A contrastive learning module is further pro-posed for identifying regular mobility patterns both temporally and across populations, allowing for a joint detection of outliers based on individual consistency and group majority patterns. Our experimental results have shown TOD4Traj's superior performance over existing models, demonstrating its effectiveness and adaptability in detecting human trajectory outliers across various datasets.
Paper Structure (23 sections, 13 equations, 3 figures, 7 tables)

This paper contains 23 sections, 13 equations, 3 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Preliminaries
  4. Methodology
  5. Modality Alignment through Spatial Temporal-to-Semantic Contrastive Learning
  6. Modeling Semantic Trajectories through Regular Pattern Contrastive Learning
  7. Quantification of Outlier Scores
  8. Experimental Results
  9. Experimental Datasets
  10. Agent-Based Simulation of Patterns of Life
  11. Real World Dataset
  12. Experimental Settings
  13. Competitor Methods
  14. Evaluation Metrics
  15. Implementation Details
  16. ...and 8 more sections

Figures (3)

  • Figure 1: A semantic trajectory of a user including location trajectory and semantic information of points of interest. For each 'check-in' location in the figure, there exists an associated text description of location information.
  • Figure 2: An example of spatial and semantic outliers. A spatial outlier for User 2 may be going to a restaurant that is unusually far away. A semantic outlier for User 1 could be going to the same courthouse that User 2 works at.
  • Figure 3: The illustration of our proposed model framework. (Top & Bottom) Extraction of spatio-temporal semantic embeddings from input trajectories; (Middle left) Contrastive learning based on human semantic trajectory periodicity; (Middle right) Quantification of trajectory anomaly scores based on dissimilarity between train and test embeddings.