pith. sign in

arxiv: 2606.17978 · v1 · pith:PO67KQBNnew · submitted 2026-06-16 · 💻 cs.AI

MoCo-AIS: A Contrastive Learning Framework for Similarity Computation of Vessel Trajectories

Ruixin Song , Md Mahbub Alam , Zahra Sadeghi , Amilcar Soares , Jos\'e F. Rodrigues-Jr , Gabriel Spadon This is my paper

Pith reviewed 2026-06-27 00:46 UTC · model grok-4.3

classification 💻 cs.AI
keywords vessel trajectorytrajectory similaritycontrastive learningMoCoAIS dataself-supervised learningtrajectory embeddingsdeep learning models
0
0 comments X

The pith

MoCo-AIS learns vessel trajectory similarities by contrasting positive and negative pairs in a self-supervised setup.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes a unified contrastive framework called MoCo-AIS that trains deep learning models to embed vessel trajectories without relying on labels from traditional distance calculations. Traditional measures carry high computational costs while supervised approaches tend to simply replicate those measures and therefore generalize poorly. By treating similarity as a task of distinguishing matched trajectory pairs from unmatched ones, the method produces embeddings that perform better on real AIS data. A sympathetic reader would care because this removes the need for expensive precomputed distances and supplies a common testbed for comparing representation models.

Core claim

MoCo-AIS adapts the Momentum Contrast paradigm to trajectory data by generating positive pairs from the same vessel track and negative pairs from different tracks, then trains an encoder to pull positives closer and push negatives apart; when applied to large real-world AIS collections that span varied navigation conditions, the resulting embeddings yield higher similarity accuracy than prior baselines while creating a consistent platform for testing multiple deep learning architectures.

What carries the argument

The MoCo contrastive setup that maintains a momentum-updated encoder and a queue of negative trajectory samples to drive representation learning from unlabeled pairs.

If this is right

  • Similarity computation for vessel trajectories improves over existing supervised and distance-based baselines on large AIS datasets.
  • A standardized benchmarking platform becomes available for comparing different deep learning models on trajectory representation tasks.
  • Self-supervised training removes dependence on large volumes of labels derived from traditional distance calculations.
  • Models trained inside the framework handle diverse navigation behaviors and operating conditions captured in real vessel tracking data.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same pair-construction logic could be applied to other mobility datasets such as road-vehicle or pedestrian tracks without redesigning the contrastive objective.
  • Once embeddings are learned, downstream tasks like route clustering or anomaly flagging could run faster because they operate on fixed-length vectors rather than raw coordinate sequences.
  • If the performance gain persists, training pipelines for mobility models could drop the preprocessing step that computes expensive pairwise distances altogether.

Load-bearing premise

That learning to separate positive and negative trajectory pairs will automatically produce embeddings whose similarity judgments generalize beyond the particular distance functions used in earlier supervised work.

What would settle it

A controlled test in which embeddings trained with MoCo-AIS are evaluated on a similarity ranking task whose ground truth comes from a distance measure never seen during contrastive training, and the learned model fails to beat a supervised baseline trained directly on that same measure.

Figures

Figures reproduced from arXiv: 2606.17978 by Amilcar Soares, Gabriel Spadon, Jos\'e F. Rodrigues-Jr, Md Mahbub Alam, Ruixin Song, Zahra Sadeghi.

Figure 1
Figure 1. Figure 1: MoCo-AIS architecture for vessel trajectory simi [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Augmentation strategies on a sample trajectory. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: RNN- and Convolutional-based encoder plugins. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Transformer-based encoder plugin. to the projected input: 𝑧˜𝑡 = 𝑓linear(𝑋𝑡 ) + P𝑡 (6) yielding a sequence 𝑧˜ ∈ R 𝑇 ×𝑑 processed by a stack of Transformer encoder layers with masking to ignore padded elements. The hid￾den states were normalized and aggregated through masked mean pooling (Eq. 7), producing a fixed-dimensional embedding. 4.2.3 Padding and Masked Mean Pooling. Each input trajectory in a batch … view at source ↗
Figure 5
Figure 5. Figure 5: Geographical coverage of three regional datasets. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: 2D projections of the learned embedding space, visu [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
read the original abstract

Trajectory similarity is a fundamental task in analyzing mobility patterns, essential for applications such as route pattern extraction, mobility prediction, and anomaly detection. Traditional distance-based measures for computing similarity incur high computational cost, driving the adoption of lightweight learning-based approaches. Supervised methods rely on extensive labels derived from traditional distance measures and often reproduce these metrics, which limits generalization. While self-supervised learning addresses this issue through contrastive learning, it lacks a unified framework, making it difficult to compare deep learning (DL) models for consistent trajectory representation. Accordingly, this paper presents MoCo-AIS, a unified framework for learning vessel trajectory embeddings based on the Momentum Contrast (MoCo) paradigm, which formulates similarity learning through positive and negative trajectory pairs. Within this framework, we evaluate a diverse set of leading DL models on large-scale, real-world vessel-tracking AIS datasets that capture diverse navigation behaviors and operating conditions. Results demonstrate that our framework significantly improves similarity learning over existing baselines, while providing a benchmarking platform for evaluating trajectory representation models.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 1 minor

Summary. The paper introduces MoCo-AIS, a unified self-supervised framework for learning vessel trajectory embeddings based on the Momentum Contrast (MoCo) paradigm. It formulates similarity computation via positive and negative trajectory pairs to overcome limitations of traditional distance measures and supervised methods that reproduce them. The framework is used to benchmark multiple deep learning models on large-scale real-world AIS datasets capturing diverse navigation behaviors, with the central claim that it significantly improves similarity learning over baselines while serving as a benchmarking platform.

Significance. If the empirical claims hold with rigorous quantitative validation, the work could establish a standardized contrastive learning approach for maritime trajectory representation, enabling better generalization in downstream tasks such as anomaly detection and route pattern extraction. The use of an established paradigm (MoCo) applied to a new domain and the explicit benchmarking platform are positive aspects that could facilitate reproducible comparisons among DL models.

minor comments (1)
  1. The abstract asserts that 'results demonstrate that our framework significantly improves similarity learning over existing baselines' but supplies no quantitative metrics, dataset sizes, model names, or baseline comparisons; the full experimental section should be checked for these details to support the claim.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for reviewing our manuscript on MoCo-AIS. We appreciate the acknowledgment of the unified self-supervised framework, the application of the MoCo paradigm to vessel trajectories, and the value of the benchmarking platform for reproducible comparisons. The recommendation of 'uncertain' appears tied to the need for rigorous validation of empirical claims; we address this below in the absence of enumerated major comments.

Circularity Check

0 steps flagged

No significant circularity; framework applies external MoCo to new domain

full rationale

The provided abstract and context describe MoCo-AIS as an application of the established external Momentum Contrast paradigm (self-supervised contrastive learning via positive/negative pairs) to vessel trajectory embeddings, followed by empirical evaluation of DL models on real-world AIS data. No equations, fitted parameters, or derivations are shown that reduce any claimed result to a quantity defined by the paper's own inputs. The benchmarking claim is presented as an empirical outcome rather than a self-referential construction, and the cited MoCo foundation is independent prior work. No self-citation load-bearing steps or self-definitional reductions appear in the given text.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no information on free parameters, axioms, or invented entities; insufficient detail available to populate the ledger.

pith-pipeline@v0.9.1-grok · 5729 in / 1033 out tokens · 33255 ms · 2026-06-27T00:46:24.240406+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

57 extracted references · 36 canonical work pages · 4 internal anchors

  1. [1]

    Md Mahbub Alam, Gabriel Spadon, Mohammad Etemad, Luis Torgo, and Evan- gelos Milios. 2024. Enhancing Short-Term Vessel Trajectory Prediction with Clustering for Heterogeneous and Multi-Modal Movement Patterns.Ocean Engineering308 (Sept. 2024), 118303. doi:10.1016/j.oceaneng.2024.118303

    work page doi:10.1016/j.oceaneng.2024.118303 2024

  2. [2]

    Md Mahbub Alam and Luis Torgo. 2022. A Clustering-based Approach for Predicting the Future Location of a Vessel.Proceedings of the Canadian Conference on Artificial Intelligence(May 2022). https://caiac.pubpub.org/pub/7rrw5ds3

    2022

  3. [3]

    Md Mahbub Alam, Luis Torgo, and Albert Bifet. 2022. A survey on spatio-temporal data analytics systems.Comput. Surveys54, 10s (2022), 1–38

    2022

  4. [4]

    Hanlin Cao, Haina Tang, Yulei Wu, Fei Wang, and Yongjun Xu. 2021. On Accurate Computation of Trajectory Similarity via Single Image Super-Resolution. In 2021 International Joint Conference on Neural Networks (IJCNN)(2021-07). 1–9. doi:10.1109/IJCNN52387.2021.9533802

    work page doi:10.1109/ijcnn52387.2021.9533802 2021

  5. [5]

    Yuan Cao, Lei Li, Xiangru Chen, Xue Xu, Zuojin Huang, and Yanwei Yu. 2024. Hypergraph Hash Learning for Efficient Trajectory Similarity Computation. InProceedings of the 33rd ACM International Conference on Information and Knowledge Management(New York, NY, USA, 2024-10-21)(CIKM ’24). Association for Computing Machinery, 175–186. doi:10.1145/3627673.3679555

    work page doi:10.1145/3627673.3679555 2024

  6. [6]

    Yanchuan Chang, Jianzhong Qi, Yuxuan Liang, and Egemen Tanin. 2023. Con- trastive Trajectory Similarity Learning with Dual-Feature Attention. In2023 IEEE 39th International Conference on Data Engineering (ICDE)(2023-04). 2933–2945. doi:10.1109/ICDE55515.2023.00224

    work page doi:10.1109/icde55515.2023.00224 2023

  7. [7]

    Jensen, and Jianzhong Qi

    Yanchuan Chang, Egemen Tanin, Gao Cong, Christian S. Jensen, and Jianzhong Qi. 2024. Trajectory Similarity Measurement: An Efficiency Perspective.Proc. VLDB Endow.17, 9 (May 2024), 2293–2306. doi:10.14778/3665844.3665858

    work page doi:10.14778/3665844.3665858 2024

  8. [8]

    Lei Chen, M Tamer Özsu, and Vincent Oria. 2005. Robust and fast similarity search for moving object trajectories. InProceedings of the 2005 ACM SIGMOD international conference on Management of data. 491–502

    2005

  9. [9]

    Yiheng Chen, Yanming Chen, Yue Cui, Xinyu Cai, Changgui Yin, and Yongxin Cheng. 2025. Optimizing Vessel Trajectories: Advanced Denoising and Inter- polation Techniques for AIS Data.Ocean Engineering327 (May 2025), 120988. doi:10.1016/j.oceaneng.2025.120988

    work page doi:10.1016/j.oceaneng.2025.120988 2025

  10. [10]

    Zhen Chen, Dalin Zhang, Shanshan Feng, Kaixuan Chen, Lisi Chen, Peng Han, and Shuo Shang. 2024. KGTS: Contrastive Trajectory Similarity Learning over Prompt Knowledge Graph Embedding. 38, 8 (2024), 8311–8319. doi:10.1609/aaai. v38i8.28672

    work page doi:10.1609/aaai 2024

  11. [11]

    Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning Phrase Representations Using RNN Encoder-Decoder for Statistical Machine Translation. arXiv:1406.1078 [cs] doi:10.48550/arXiv.1406.1078

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1406.1078 2014

  12. [12]

    David H Douglas and Thomas K Peucker. 1973. Algorithms for the Reduction of the Number of Points Required to Represent a Digitized Line or Its Caricature. Cartographica10, 2 (Dec. 1973), 112–122. doi:10.3138/FM57-6770-U75U-7727

    work page doi:10.3138/fm57-6770-u75u-7727 1973

  13. [13]

    Martha Dais Ferreira, Gabriel Spadon, Amilcar Soares, and Stan Matwin. 2022. A Semi-Supervised Methodology for Fishing Activity Detection Using the Geometry behind the Trajectory of Multiple Vessels.Sensors (Basel, Switzerland)22, 16 (Aug. 2022), 6063. doi:10.3390/s22166063

    work page doi:10.3390/s22166063 2022

  14. [14]

    Carlos Andres Ferrero, Luis Otavio Alvares, Willian Zalewski, and Vania Bogorny

  15. [15]

    InProceedings of the 33rd Annual ACM Symposium on Applied Computing (SAC ’18)

    MOVELETS: Exploring Relevant Subtrajectories for Robust Trajectory Classification. InProceedings of the 33rd Annual ACM Symposium on Applied Computing (SAC ’18). Association for Computing Machinery, New York, NY, USA, 849–856. doi:10.1145/3167132.3167225

    work page doi:10.1145/3167132.3167225

  16. [16]

    Carlos Andres Ferrero, Lucas May Petry, Luis Otavio Alvares, Camila Leite da Silva, Willian Zalewski, and Vania Bogorny. 2020. MasterMovelets: Discovering Heterogeneous Movelets for Multiple Aspect Trajectory Classification.Data Mining and Knowledge Discovery34, 3 (May 2020), 652–680. doi:10.1007/s10618- 020-00676-x

    work page doi:10.1007/s10618- 2020

  17. [17]

    Mélanie Fournier, R Casey Hilliard, Sara Rezaee, and Ronald Pelot. 2018. Past, present, and future of the satellite-based automatic identification system: Areas of applications (2004–2016).WMU journal of maritime affairs17, 3 (2018), 311–345

    2018

  18. [18]

    Andre Salvaro Furtado, Despina Kopanaki, Luis Otavio Alvares, and Vania Bo- gorny. 2016. Multidimensional Similarity Measuring for Semantic Trajectories. Transactions in GIS20, 2 (2016), 280–298. doi:10.1111/tgis.12156

    work page doi:10.1111/tgis.12156 2016

  19. [19]

    Richemond, Elena Buchatskaya, Carl Doersch, Bernardo Avila Pires, Zhao- han Daniel Guo, Mohammad Gheshlaghi Azar, Bilal Piot, Koray Kavukcuoglu, Rémi Munos, and Michal Valko

    Jean-Bastien Grill, Florian Strub, Florent Altché, Corentin Tallec, Pierre H. Richemond, Elena Buchatskaya, Carl Doersch, Bernardo Avila Pires, Zhao- han Daniel Guo, Mohammad Gheshlaghi Azar, Bilal Piot, Koray Kavukcuoglu, Rémi Munos, and Michal Valko. 2020. Bootstrap Your Own Latent a New Ap- proach to Self-Supervised Learning. InProceedings of the 34th ...

    2020

  20. [20]

    1914.Grundzüge der Mengenlehre

    Felix Hausdorff. 1914.Grundzüge der Mengenlehre. Veit, Leipzig. In- troduces the Hausdorff metric (Pompeiu–Hausdorff distance). English trans- lation:Set Theory, Chelsea Publishing, 1957.. https://archive.org/details/ grundzgedermen00hausuoft

    1914

  21. [21]

    2020.Momentum Contrast for Unsupervised Visual Representation Learning

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. 2020.Momentum Contrast for Unsupervised Visual Representation Learning. arXiv:1911.05722 [cs] doi:10.48550/arXiv.1911.05722

    work page doi:10.48550/arxiv.1911.05722 2020

  22. [22]

    Liang Huang, Chengpeng Wan, Yuanqiao Wen, Rongxin Song, and Pieter van Gelder. 2024. Generation and application of maritime route networks: overview and future research directions.IEEE Transactions on Intelligent Transportation Systems(2024)

    2024

  23. [23]

    Eamonn Keogh and Chotirat Ann Ratanamahatana. 2005. Exact Indexing of Dynamic Time Warping.Knowledge and Information Systems7, 3 (March 2005), 358–386. doi:10.1007/s10115-004-0154-9

    work page doi:10.1007/s10115-004-0154-9 2005

  24. [24]

    Colin Lea, Rene Vidal, Austin Reiter, and Gregory D. Hager. 2016. Tempo- ral Convolutional Networks: A Unified Approach to Action Segmentation. arXiv:1608.08242 [cs] doi:10.48550/arXiv.1608.08242

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1608.08242 2016

  25. [25]

    HiT-JEPA: A Hierarchical Self-supervised Trajectory Embedding Framework for Similarity Computation

    Lihuan Li, Hao Xue, Shuang Ao, Yang Song, and Flora Salim. 2025.HiT-JEPA: A Hierarchical Self-supervised Trajectory Embedding Framework for Similarity Computation. arXiv:2507.00028 [cs] doi:10.48550/arXiv.2507.00028

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2507.00028 2025

  26. [26]

    Jensen, and Wei Wei

    Xiucheng Li, Kaiqi Zhao, Gao Cong, Christian S. Jensen, and Wei Wei. 2018. Deep Representation Learning for Trajectory Similarity Computation. In2018 IEEE 34th International Conference on Data Engineering (ICDE)(2018-04). 617–628. doi:10.1109/ICDE.2018.00062

    work page doi:10.1109/icde.2018.00062 2018

  27. [27]

    Maohan Liang, Ryan Wen Liu, Ruobin Gao, Zhe Xiao, Xiaocai Zhang, and Hua Wang. 2024. A Survey of Distance-Based Vessel Trajectory Clustering: Data Pre-processing, Methodologies, Applications, and Experimental Evaluation. arXiv:2407.11084 [eess] doi:10.48550/arXiv.2407.11084

    work page doi:10.48550/arxiv.2407.11084 2024

  28. [28]

    Maohan Liang, Ryan Wen Liu, Shichen Li, Zhe Xiao, Xin Liu, and Feng Lu. 2021. An Unsupervised Learning Method with Convolutional Auto-Encoder for Vessel Trajectory Similarity Computation.Ocean Engineering225 (April 2021), 108803. arXiv:2101.03169 [cs] doi:10.1016/j.oceaneng.2021.108803

    work page doi:10.1016/j.oceaneng.2021.108803 2021

  29. [29]

    Tiantian Liu, Hengyu Liu, Tianyi Li, Kristian Torp, and Christian S. Jensen. 2025. ACTIVE: Continuous Similarity Search for Vessel Trajectories. arXiv:2504.01142 [cs] doi:10.48550/arXiv.2504.01142

    work page doi:10.48550/arxiv.2504.01142 2025

  30. [30]

    Xiang Liu, Xiaoying Tan, Yuchun Guo, Yishuai Chen, and Zhe Zhang. 2022. CSTRM: Contrastive Self-Supervised Trajectory Representation Model for Tra- jectory Similarity Computation. 185 (2022), 159–167. doi:10.1016/j.comcom.2022. 01.001

    work page doi:10.1016/j.comcom.2022 2022

  31. [31]

    Sizhe Luo and Weiming Zeng. 2023. Vessel Trajectory Similarity Computation Based on Heterogeneous Graph Neural Network. 11, 7 (2023), 1318. Issue 7. doi:10.3390/jmse11071318

    work page doi:10.3390/jmse11071318 2023

  32. [32]

    Pin Nie, Zhenjie Chen, Nan Xia, Qiuhao Huang, and Feixue Li. 2021. Trajectory Similarity Analysis with the Weight of Direction and K-Neighborhood for AIS Data.ISPRS International Journal of Geo-Information10, 11 (Nov. 2021), 757. doi:10.3390/ijgi10110757

    work page doi:10.3390/ijgi10110757 2021

  33. [33]

    Zahra Sadeghi and Stan Matwin. 2023. Anomaly Detection for Maritime Naviga- tion Based on Probability Density Function of Error of Reconstruction.Journal of Intelligent Systems32, 1 (Jan. 2023). doi:10.1515/jisys-2022-0270

    work page doi:10.1515/jisys-2022-0270 2023

  34. [34]

    Schuster and K.K

    M. Schuster and K.K. Paliwal. 1997. Bidirectional Recurrent Neural Networks. IEEE Transactions on Signal Processing45, 11 (Nov. 1997), 2673–2681. doi:10.1109/ 78.650093

    1997

  35. [35]

    2025.Towards Robust Trajectory Embedding for Similarity Computation: When Triangle Inequality Violations in Distance Metrics Matter

    Jianing Si, Haitao Yuan, Nan Jiang, Minxiao Chen, Xiao Ma, and Shang- guang Wang. 2025.Towards Robust Trajectory Embedding for Similarity Computation: When Triangle Inequality Violations in Distance Metrics Matter. arXiv:2504.10933 [cs] doi:10.48550/arXiv.2504.10933

    work page doi:10.48550/arxiv.2504.10933 2025

  36. [36]

    Ana Paula Ramos Souza, Chiara Renso, Raffaele Perego, and Vania Bogorny. 2022. MAT-Index: An Index for Fast Multiple Aspect Trajectory Similarity Measuring. Transactions in GIS26, 2 (2022), 691–716. doi:10.1111/tgis.12889

    work page doi:10.1111/tgis.12889 2022

  37. [37]

    Gabriel Spadon, Andre CPLF de Carvalho, Jose F Rodrigues-Jr, and Luiz GA Alves

  38. [38]

    Reconstructing commuters network using machine learning and urban indicators.Scientific reports9, 1 (2019), 11801

    2019

  39. [39]

    Han Su, Shuncheng Liu, Bolong Zheng, Xiaofang Zhou, and Kai Zheng. 2020. A survey of trajectory distance measures and performance evaluation.The VLDB Journal29, 1 (2020), 3–32

    2020

  40. [40]

    Yongfeng Suo, Wenke Chen, Christophe Claramunt, and Shenhua Yang. 2020. A ship trajectory prediction framework based on a recurrent neural network. Sensors20, 18 (2020), 5133

    2020

  41. [41]

    Yaguang Tao, Alan Both, Rodrigo I Silveira, Kevin Buchin, Stef Sijben, Ross S Purves, Patrick Laube, Dongliang Peng, Kevin Toohey, and Matt Duckham. 2021. A comparative analysis of trajectory similarity measures.GIScience & Remote Sensing58, 5 (2021), 643–669

    2021

  42. [42]

    Tarlis Tortelli Portela, Jonata Tyska Carvalho, and Vania Bogorny. 2022. HiPer- Movelets: High-Performance Movelet Extraction for Trajectory Classification. International Journal of Geographical Information Science36, 5 (May 2022), 1012–

    2022

  43. [43]

    doi:10.1080/13658816.2021.2018593

    work page doi:10.1080/13658816.2021.2018593 2021

  44. [44]

    Alexandros Troupiotis-Kapeliaris, Georgios Grigoropoulos, Marios Vodas, and Konstantina Bereta. 2025. Dynamic Weather-Resilient Vessel Routing Using Big AIS Data. InProceedings of the 33rd ACM International Conference on Advances in Geographic Information Systems. Association for Computing Machinery, New York, NY, USA, 706–716

    2025

  45. [45]

    Alexandros Troupiotis-Kapeliaris, Christos Kastrisios, and Dimitris Zissis. 2025. Vessel Trajectory Data Mining: A Review.IEEE Access(2025). MoCo-AIS: Contrastive Learning for Vessel Trajectory Similarity Conference acronym ’XX, June 03–05, 2018, Woodstock, NY

    2025

  46. [46]

    Enmei Tu, Guanghao Zhang, Lily Rachmawati, Eshan Rajabally, and Guang- Bin Huang. 2018. Exploiting AIS Data for Intelligent Maritime Navigation: A Comprehensive Survey From Data to Methodology.IEEE Transactions on Intelligent Transportation Systems19, 5 (May 2018), 1559–1582. doi:10.1109/TITS. 2017.2724551

    work page doi:10.1109/tits 2018

  47. [47]

    Aaron Van den Oord, Yazhe Li, and Oriol Vinyals. 2019. Representation Learning with Contrastive Predictive Coding. arXiv:1807.03748 [cs] doi:10.48550/arXiv. 1807.03748

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv 2019

  48. [48]

    Iraklis Varlamis, Christos Sardianos, Vania Bogorny, Luis Otavio Alvares, Jô- nata Tyska Carvalho, Chiara Renso, Raffaele Perego, and John Violos. 2021. A Novel Similarity Measure for Multiple Aspect Trajectory Clustering. In Proceedings of the 36th Annual ACM Symposium on Applied Computing (SAC ’21). Association for Computing Machinery, New York, NY, USA...

    work page doi:10.1145/3412841.3441935 2021

  49. [49]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention Is All You Need. InAdvances in Neural Information Processing Systems, I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett (Eds.), Vol. 30. Curran Associates, Inc

    2017

  50. [50]

    Weinberger and Lawrence K

    Kilian Q. Weinberger and Lawrence K. Saul. 2009. Distance Metric Learning for Large Margin Nearest Neighbor Classification.Journal of Machine Learning Research10, 9 (2009), 207–244

    2009

  51. [51]

    Larissa Westerdijk, M Beekveld, AE Eiben, and EN Belitser. 2019. Classifying vessel types based on AIS data.MSc, Vrije University, Amsterdam, Holland(2019), 49–61

    2019

  52. [52]

    Peilun Yang, Hanchen Wang, Ying Zhang, Lu Qin, Wenjie Zhang, and Xuemin Lin. 2021. T3S: Effective Representation Learning for Trajectory Similarity Computation. In2021 IEEE 37th International Conference on Data Engineering (ICDE)(2021-04). 2183–2188. doi:10.1109/ICDE51399.2021.00221

    work page doi:10.1109/icde51399.2021.00221 2021

  53. [53]

    Di Yao, Gao Cong, Chao Zhang, and Jingping Bi. 2019. Computing Trajectory Similarity in Linear Time: A Generic Seed-Guided Neural Metric Learning Ap- proach. In2019 IEEE 35th International Conference on Data Engineering (ICDE). 1358–1369. doi:10.1109/ICDE.2019.00123

    work page doi:10.1109/icde.2019.00123 2019

  54. [54]

    Di Yao, Haonan Hu, Lun Du, Gao Cong, Shi Han, and Jingping Bi. 2022. Traj- GAT: A Graph-based Long-term Dependency Modeling Approach for Trajectory Similarity Computation. InProceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining(New York, NY, USA, 2022-08-14)(KDD ’22). Association for Computing Machinery, 2275–2285. doi:10.11...

    work page doi:10.1145/3534678.3539358 2022

  55. [55]

    Hanyuan Zhang, Xinyu Zhang, Qize Jiang, Baihua Zheng, Zhenbang Sun, Wei- wei Sun, and Changhu Wang. 2020. Trajectory Similarity Learning with Aux- iliary Supervision and Optimal Matching. InProceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence. International Joint Con- ferences on Artificial Intelligence Organization, Y...

    work page doi:10.24963/ijcai.2020/444 2020

  56. [56]

    Liye Zhang, Qiang Meng, Zhe Xiao, and Xiuju Fu. 2018. A Novel Ship Trajectory Reconstruction Approach Using AIS Data.Ocean Engineering159 (July 2018), 165–174. doi:10.1016/j.oceaneng.2018.03.085

    work page doi:10.1016/j.oceaneng.2018.03.085 2018

  57. [57]

    Yu Zheng. 2015. Trajectory data mining: an overview.ACM Transactions on Intelligent Systems and Technology (TIST)6, 3 (2015), 1–41. Received 20 February 2007; revised 12 March 2009; accepted 5 June 2009

    2015