Pith. sign in

REVIEW 4 major objections 6 minor 33 references

A Framework for Mining Collectively-Behaving Bots in MMORPGs

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read An unsupervised framework, BotTRep, mines collectively-behaving bot groups in MMORPGs from unlabeled trajectory data by learning context-aware embeddings and clustering them with DBSCAN, and the paper shows it outperforms autoencoder…

desk verdict A solid applied framework for mining bot groups from unlabeled MMORPG trajectories, but the central 'high accuracy' claim outruns the evidence because the main metric is partly circular and there is no ground-truth validation. read the letter →

arxiv 2501.10461 v2 pith:L6GEPZA3 submitted 2025-01-15 cs.LG cs.AI

Hyunsoo Kim , Jun Hee Kim , Jaeman Son , Jihoon Song , Eunjo Lee This is my paper · ORCID
classification cs.LGcs.AI
keywords gamebotdetectionMMORPGtrajectoryrepresentationlearningcontrastiveDBSCANunsupervisedcollectivebehavioraccessinformationhomogeneity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

BotTRep is an unsupervised framework for mining collectively-behaving bot groups in MMORPGs from completely unlabeled in-game trajectory data. It learns to embed each player's daily movement so that contextually similar trajectories end up close together, then applies DBSCAN to isolate dense clusters as suspected bot groups. The paper claims this is the first trajectory-based bot mining approach that is both label-free and practical for industrial service, training in about 8.5 hours so that monitoring can restart after daily game patches. If correct, it lets game companies detect coordinated bot groups without a labeling infrastructure and with visual evidence that can support legal ban decisions.

What carries the argument

The central machinery is the BotTRep encoder, a Transformer with contrastive triplets. Each minute-level location is tokenized into a zone (256-by-256 coordinate block) and a cell (8-by-8 block), and each token embeds as the sum of a zone embedding, a cell embedding, and a timestamp encoding; this design intentionally replaces raw Euclidean distance with contextual region co-occurrence, so teleportation between a village and a hunting ground is treated as 'close' when players regularly visit both. The encoder is trained jointly with triplet margin loss (pulling anchor and positive trajectory segments together while pushing a random negative apart) and masked cell prediction (predicting masked cell tokens in the anchor), which together force the model to capture spatiotemporal context. During inference, entire daily trajectories are embedded and clustered with DBSCAN, using an epsilon chosen from the 4-nearest-neighbor distance distribution; all non-noise clusters are defined as collectively-behaving groups. The design addresses the two MMORPG-specific problems identified in the paper: teleportation breaking spatial-distance assumptions, and multiple instance dungeons sharing the same local coordinate system.

What would settle it

A direct falsifier would be to obtain a set of accounts later confirmed as bots (e.g., from bans or real-money-trading investigations) and check whether they fall disproportionately inside the mined clusters. If confirmed bot accounts appear as often in the noise class as inside clusters, or if clusters contain many legitimate players with coincidentally overlapping routes, the central claim would fail.

Watch

Extended reading notes

Core claim

The paper's central claim is that contrastive trajectory representation learning followed by DBSCAN clustering accurately mines collectively-behaving bot groups from raw, unlabeled MMORPG movement logs. On the authors' own evaluation, the mined clusters show a time-aware Jaccard similarity of about 0.36 against under 0.001 for random pairs, and an access-information homogeneity of 1.0079, meaning that players inside a cluster share the same IP or device network—the signature of a single operator controlling multiple avatars. These numbers beat three autoencoder baselines (Bi-GRU, Bi-LSTM, and a plain Transformer) on both metrics, while completing training in 8.5 hours versus more than 24 hours for the baselines. The authors therefore conclude that the framework has effectively detected the bot groups it aimed to identify, and that the clusters are genuinely coordinated rather than coincidental overlaps.

Load-bearing premise

The load-bearing premise is that a DBSCAN cluster of players with contextually similar trajectories and overlapping access information is actually a group of bots, an assumption never validated against a gold standard of known bot accounts or bans.

Editorial extensions

If this is right

  • Game companies can deploy BotTRep without a labeling pipeline, since the framework mines bot groups from unlabeled movement logs alone.
  • The 8.5-hour training time fits the paper's stated service constraint, allowing daily monitoring to start at 9 AM after each game patch and to finish within the 9-hour window.
  • The visualization heatmap gives game masters a human-checkable record of each cluster's synchronized movement, which the paper argues lowers false-banning risk and provides evidence if a ban is challenged.
  • The access-information homogeneity of about 1.0 implies that players inside the same mined cluster are typically operated by the same person or network, supporting the interpretation that these are real bot farms.
  • The two-task training (contrastive triplet loss plus masked cell prediction) yields better contextual similarity and homogeneity than autoencoder-based baselines, suggesting direct similarity supervision is more effective than reconstruction for trajectory representation in this domain.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension is to run BotTRep on a game that lacks access-information logs; if clusters still show high time-aware Jaccard similarity, the method would generalize to environments without IP or device linkage data.
  • The contextual co-occurrence principle suggests a broader design rule for trajectory representation: when movement includes long-range jumps or portals, semantic region proximity matters more than raw spatial distance, a lesson that could transfer to delivery fleets, transit networks, or any domain with discontinuous movement.
  • The paper evaluates clusters against proxies rather than a true bot gold standard; augmenting the framework with a small set of confirmed banned accounts as a validation anchor would turn the proxy metrics into calibrated precision and recall.
  • The definition of collectively-behaving groups as 'all non-noise DBSCAN clusters' is a strong prior; an inference worth testing is whether clusters that persist across multiple consecutive days correspond to the same underlying bot operators, which would strengthen the claim that the clusters are stable entities.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes BotTRep, an unsupervised framework for mining collectively-behaving bot groups in MMORPGs. Trajectories are binned into zone and cell tokens, embedded with a Transformer trained jointly on triplet margin contrastive learning and masked cell prediction, and then clustered with DBSCAN. The method is evaluated on seven days of unlabeled Lineage W trajectories using two metrics: time-aware Jaccard contextual similarity within detected clusters versus random pairs, and access-information homogeneity across cluster members. The authors report that BotTRep achieves the best contextual similarity and lowest homogeneity value, and conclude that it detected the targeted bot groups with high accuracy.

Significance. If its central claim were properly supported, BotTRep would be a valuable industrial contribution: it addresses label scarcity, long-sequence training constraints, and explainability for game-master review, and it is evaluated on real production data with a useful ablation study. The explicit treatment of the nine-hour training deadline, the lightweight contrastive design, and the visualization tool are genuine strengths. However, the manuscript's load-bearing assertion of "high accuracy" is not supported by the evidence presented, because the evaluation lacks ground-truth bot labels and the primary metric is partly circular with the clustering itself. The paper is an interesting system description, but it needs external validation or substantially tempered claims before it can be accepted as a scientific claim about bot detection accuracy.

major comments (4)
  1. [Section 3.5 and Section 4.2 (Eqs. 9-11)] The evaluation is partially circular. In Section 3.5 the authors state 'we decided to define all the clustered groups (i.e. DBSCAN did not classify as noise) as collectively-behaving groups,' and in Section 4.2 positive pairs are constructed from those same cluster labels via Eq. (9), with contextual similarity measured as time-aware Jaccard similarity of raw trajectories. Since DBSCAN groups trajectories whose learned embeddings are close, and the model was trained to make contextually similar trajectories closer, high within-cluster contextual similarity is an internal-consistency check rather than evidence that the members are bots. To support the claim that the framework 'effectively detected the bot groups we aimed to identify with high accuracy' (Section 4.4), the authors need an external outcome measure: precision/recall against known bot accounts or banned-player lists, a manual audit by game masters with stated criteria, or at minimum a comparison against a random-clustering baseline to show that the observed similarity is not a trivial artifact of clustering.
  2. [Section 4.2 (Access information homogeneity)] Access-information homogeneity is independent of the learned representation, which is a strength, but it cannot by itself certify that a cluster is a bot group. Identical IP, device, or network information is also consistent with benign multi-accounting, shared households, internet cafés, or guild members using a common VPN. The reported value of 1.0079 in Table 1 is suggestive, but without a base-rate comparison for randomly formed groups or for known benign clusters, it does not establish that the clusters are automated bot groups. The sentence in Section 4.4 that the access information of all players in a cluster is 'related, signifying our model has effectively detected the bot groups' therefore overstates what the metric can show.
  3. [Section 4.3 and Table 1] The clustering parameter q (the epsilon quantile) is selected on the same downstream dataset used to report final results. The rows labelled (c) in Table 1 vary q and the paper chooses q=0.05 because it yields the best contextual similarity and lowest access-information homogeneity; this is test-set tuning and can inflate apparent performance. The authors should use a separate validation period or a nested selection procedure for q and for the other hyperparameters, or report all settings without claiming the best one as the final model.
  4. [Section 4.3 and Section 4.5] The only evidence connecting clusters to actual bot behavior is anecdotal: the YouTube example in Section 4.3 (average Jaccard around 0.3, access homogeneity 1.0) and the heatmap visualizations in Fig. 5. No systematic manual audit of detected clusters is reported, and the paper does not provide any measure that can detect false positives, such as the proportion of clusters later confirmed by game masters or the proportion of cluster members with no automation evidence. Without such an outcome, the central claim remains an assertion rather than a demonstrated result.
minor comments (6)
  1. [Section 3.2] The heading 'Traning dataset' contains a typo and should be 'Training dataset'.
  2. [Section 2.1] The game name is misspelled as 'Linage W' in the paragraph on local coordinate systems; it should be 'Lineage W'.
  3. [Fig. 4 caption] The caption says 'The thrid image illustrates clustering results'; 'thrid' should be 'third'.
  4. [Eq. (7) and surrounding text] The notation in Eq. (7) could be clarified: it applies Linear(·) to f^A_j for j in δ, but the text first says 'among these 16 extracted results, Linear(·) is applied to the tokens that had been masked'; make explicit that j indexes positions in the anchor sequence, not token identities.
  5. [Table 2] The 'Training time Minutes (per epoch)' column is awkwardly formatted; consider splitting into total training time and minutes per epoch, and indicate the hardware configuration in the caption.
  6. [Section 4.3] The YouTube link is not a citable evaluation artifact; if it is intended to support the claim about detected bot examples, it should be replaced with supplementary material or a persistent repository.

Circularity Check

2 steps flagged · score 5.0 of 10

The 'high accuracy' bot-group claim is supported mainly by metrics built from the detector's own clusters; the target class is defined as the DBSCAN output, and the primary similarity metric reuses those same clusters as positive pairs.

  1. self definitional [Section 3.5 (Clustering collectively-behaving groups) and Section 4.3 (Experimental results)]
    "Hence, we decided to define all the clustered groups (i.e. DBSCAN did not classify as noise) as collectively-behaving groups. ... This indicates that the access information of all players within a cluster is related, signifying our model has effectively detected the bot groups we aimed to identify with high accuracy."

    The paper defines the target class ('collectively-behaving groups') as the DBSCAN clusters produced by its own framework, then concludes that the model 'effectively detected the bot groups we aimed to identify with high accuracy.' With the target defined as the cluster output, the detection result is a restatement of the operational definition rather than a comparison against independently known bot accounts. No ground-truth bot labels or banned-player validation are used, so the only support for the accuracy claim is the definitional equivalence plus proxy metrics.

  2. other [Section 4.2, Eqs. (9)-(11)]
    "we generated positive and negative groups based on their representations using DBSCAN. Then, we excluded noise-labeled data determined by the DBSCAN algorithm. We then selected positive and negative pairs based on the cluster labels in this way: pos = {(T pi , T pj ) | i = 1, .., N, j∈ ξ(T ′pi )}"

    The headline contextual-similarity metric takes its positive pairs from the DBSCAN clusters, and those clusters are fitted from representation vectors the model was explicitly trained to make closer for contextually similar trajectories. Measuring time-aware Jaccard similarity within these same clusters is therefore an internal-consistency check of the clustering, not an external test of whether the clusters are bots. The access-information homogeneity is an independent signal, but it is an indirect proxy (shared IP/device/network can also occur for benign multi-account or shared-network players) and is not compared to known bot labels.

full rationale

The derivation chain is partially circular. Section 3.5 makes the system's own DBSCAN clusters the operational definition of the collectively-behaving groups the paper aims to mine, and the final accuracy claim is then justified by properties of those same clusters. Section 4.2's primary metric compounds this by constructing positive pairs from the cluster labels produced by the representation-clustering pipeline, which the training objective was designed to make similar; the metric therefore largely validates the clustering against its own training criterion. The access-information homogeneity metric is genuinely independent of the representation and gives some non-circular evidence, but it is still a proxy rather than a ground-truth bot label, and the paper offers no precision/recall against known or manually audited bot accounts. The self-citations in the paper are not load-bearing and do not contribute to the circularity. Overall, the central 'high accuracy' verdict is not forced entirely by definition because the access-information signal and the visualization provide separable evidence, so the circularity is partial rather than total.

Assumptions & free parameters 9 free parameters · 6 assumptions · 0 invented entities

No new physical entities are introduced. The main costs are the many hyperparameters chosen empirically on the proprietary dataset, and the domain assumptions that equate trajectory clusters with bot behavior without a ground-truth validation.

free parameters (9)
  • zone size = 256 x 256 coordinates
    Chosen empirically; authors recommend half the size of an instance dungeon to avoid over-abstraction (Section 3.1).
  • cell size = 8 x 8 coordinates
    Chosen empirically; authors use approximately the range of a ranged character to detect archer/mage movement (Section 3.1).
  • training sequence length and triplet length = 32 for input, 16 for anchor/positive/negative
    Set to balance information content and distinguishability; longer sequences become too similar across positive and negative pairs (Section 3.2).
  • masking probability r = 0.2 (best) vs 0.3
    Masking ratio for the MCP task; r=0.2 chosen from ablation results (Section 4.3, Table 1).
  • embedding dimension d_model = 256 (best) vs 512
    Model width; tuned in ablation study (Section 4.3, Table 1).
  • inner dimension d_hid = 1024 (best) vs 2048
    Transformer feed-forward dimension; tuned in ablation study (Section 4.3, Table 1).
  • triplet margin beta = 0.5 (best) vs 1.0
    Margin in triplet loss; tuned in ablation study (Section 4.3, Table 1).
  • DBSCAN min_samples = 4
    Set to 4 because the paper targets suspicious groups with more than 4 players, an industrial requirement (Section 3.5).
  • DBSCAN epsilon quantile q = 0.05 (best) vs 0.10, 0.15, 0.20
    Epsilon is chosen as a quantile of k-NN distances; q=0.05 maximizes contextual similarity and is selected from the ablation table without a separate validation set (Section 4.3, Table 1).
assumptions (6)
  • domain assumption Bots tend to behave collectively with similar in-game trajectories due to auto-programs
    Core motivation stated in the abstract and Section 1.
  • domain assumption Groups of 4 or more players exhibiting contextually similar trajectories are collectively-behaving bots
    Definition in Sections 3.1 and 3.5: 'we decided to define all the clustered groups ... as collectively-behaving groups.'
  • domain assumption DBSCAN noise labels correspond to benign players
    Section 3.5: 'Such points generally correspond to benign players because they have peculiar trajectories resulting from diverse preferences in play styles.'
  • domain assumption Access-information homogeneity indicates a single operator controlling multiple accounts
    Section 4.2: 'if the access information of players mined as collectively-behaving groups is identical, it signifies that they are indeed real collectively-behaving bots controlled by the same user.'
  • ad hoc to paper Zone and cell binning at 256 and 8 units preserves the spatial context needed to distinguish trajectories
    Zone/cell sizes are set empirically in Section 3.1 with heuristics about dungeon sizes and ranged attacks; no independent theoretical justification is provided.
  • domain assumption Triplet contrastive loss with random timestamp encoding yields robust trajectory representations
    Design choice in Section 3.3, justified by alignment/uniformity arguments from [26] and the need for fast training, but not independently validated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Framework for Mining Collectively-Behaving Bots in MMORPGs." pith.science (2026). https://pith.science/paper/L6GEPZA3

@misc{pith2026250110461,
  author       = {Pith},
  title        = {Pith review of: A Framework for Mining Collectively-Behaving Bots in MMORPGs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L6GEPZA3}},
  note         = {Machine review of arXiv:2501.10461}
}
read the original abstract

In MMORPGs (Massively Multiplayer Online Role-Playing Games), abnormal players (bots) using unauthorized automated programs to carry out pre-defined behaviors systematically and repeatedly are commonly observed. Bots usually engage in these activities to gain in-game money, which they eventually trade for real money outside the game. Such abusive activities negatively impact the in-game experiences of legitimate users since bots monopolize specific hunting areas and obtain valuable items. Thus, detecting abnormal players is a significant task for game companies. Motivated by the fact that bots tend to behave collectively with similar in-game trajectories due to the auto-programs, we developed BotTRep, a framework that comprises trajectory representation learning followed by clustering using a completely unlabeled in-game trajectory dataset. Our model aims to learn representations for in-game trajectory sequences so that players with contextually similar trajectories have closer embeddings. Then, by applying DBSCAN to these representations and visualizing the corresponding moving patterns, our framework ultimately assists game masters in identifying and banning bots.

Figures

Figures reproduced from arXiv: 2501.10461 by the authors.

Figure 1
Figure 1. Spatially, entities (a) and (b) are in close proximity; however, if the overall context between (b) and (c) is similar, then the embeddings of (b) and (c) should be closer [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. The following subsections provide a detailed explanation of the design [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. The first image shows a comparison between the [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: This image shows player locations over time based on clustering results. [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 25 canonical work pages

  1. [1]

    In: Entertainment Computing-ICEC 2008: 7th International Conference, A Framework for Mining Collectively-Behaving Bots in MMORPGs 17 Pittsburgh, PA, USA, September 25-27, 2008

    Chen, K.T., Liao, A., Pao, H.K.K., Chu, H.H.: Game bot detection based on avatar trajectory. In: Entertainment Computing-ICEC 2008: 7th International Conference, A Framework for Mining Collectively-Behaving Bots in MMORPGs 17 Pittsburgh, PA, USA, September 25-27, 2008. Proceedings 7. pp. 94–105. Springer (2009)

    work page 2009

  2. [2]

    In: Proceedings of the Thirtieth international conference on Very large data bases-Volume 30

    Chen, L., Ng, R.: On the marriage of lp-norms and edit distance. In: Proceedings of the Thirtieth international conference on Very large data bases-Volume 30. pp. 792–803 (2004)

    work page 2004

  3. [3]

    In: Proceedings of the 2005 ACM SIGMOD international conference on Management of data

    Chen, L., Özsu, M.T., Oria, V.: Robust and fast similarity search for moving object trajectories. In: Proceedings of the 2005 ACM SIGMOD international conference on Management of data. pp. 491–502 (2005)

    work page 2005

  4. [4]

    In: 2011 IEEE 27th International Conference on Data Engineering

    Chen, Z., Shen, H.T., Zhou, X.: Discovering popular routes from trajectories. In: 2011 IEEE 27th International Conference on Data Engineering. pp. 900–911. IEEE (2011)

    work page 2011

  5. [5]

    Ester, M., Kriegel, H.P., Sander, J., Xu, X., et al.: A density-based algorithm for discovering clusters in large spatial databases with noise. In: kdd. vol. 96, pp. 226–231 (1996)

    1996

  6. [6]

    In: Proceedings of the AAAI Conference on Artificial Intelligence

    Feng, S., Cong, G., An, B., Chee, Y.M.: Poi2vec: Geographical latent representation for predicting future visitors. In: Proceedings of the AAAI Conference on Artificial Intelligence. vol. 31 (2017)

    work page 2017

  7. [7]

    IEEE Access 4, 2056–2067 (2016)

    Feng, Z., Zhu, Y.: A survey on trajectory data mining: Techniques and applications. IEEE Access 4, 2056–2067 (2016)

    work page 2016

  8. [8]

    Available at SSRN 1089307 (2008)

    Huhh, J.S.: Simple economics of real-money trading in online games. Available at SSRN 1089307 (2008)

    work page 2008

Show all 33 references
  1. [9]

    Jaccard, P.: The distribution of the flora in the alpine zone. 1. New phytologist 11(2), 37–50 (1912)

    1912

  2. [10]

    arXiv preprint arXiv:2312.16015 (2023)

    Jadon, A., Patil, A.: A comprehensive survey of evaluation techniques for recom- mendation systems. arXiv preprint arXiv:2312.16015 (2023)

    2023 arXiv

  3. [11]

    In: Proceedings of naacL-HLT

    Kenton, J.D.M.W.C., Toutanova, L.K.: Bert: Pre-training of deep bidirectional transformers for language understanding. In: Proceedings of naacL-HLT. vol. 1, p. 2 (2019)

    2019

  4. [12]

    In: Proceedings of the 2018 World Wide Web Conference

    Lee, E., Woo, J., Kim, H., Kim, H.K.: No silk road for online gamers! using social network analysis to unveil black markets in online games. In: Proceedings of the 2018 World Wide Web Conference. pp. 1825–1834 (2018)

    2018

  5. [13]

    In: 2018 IEEE 34th international conference on data engineering (ICDE)

    Li, X., Zhao, K., Cong, G., Jensen, C.S., Wei, W.: Deep representation learning for trajectory similarity computation. In: 2018 IEEE 34th international conference on data engineering (ICDE). pp. 617–628. IEEE (2018)

    2018

  6. [14]

    Journal of machine learning research 9(11) (2008)

    Van der Maaten, L., Hinton, G.: Visualizing data using t-sne. Journal of machine learning research 9(11) (2008)

    2008

  7. [15]

    IEEE Transactions on Computational Intelligence and AI in Games2(3), 162–175 (2010)

    Pao, H.K., Chen, K.T., Chang, H.C.: Game bot detection via avatar trajectory analysis. IEEE Transactions on Computational Intelligence and AI in Games2(3), 162–175 (2010)

    2010

  8. [16]

    In: Advances in Knowledge Discovery and Data Mining: 26th Pacific- Asia Conference, PAKDD 2022, Chengdu, China, May 16–19, 2022, Proceedings, Part II

    Qi, X., Pu, J., Zhao, S., Wu, R., Tao, J.: A gnn-enhanced game bot detection model for mmorpgs. In: Advances in Knowledge Discovery and Data Mining: 26th Pacific- Asia Conference, PAKDD 2022, Chengdu, China, May 16–19, 2022, Proceedings, Part II. pp. 316–327. Springer (2022)

    2022

  9. [17]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Schroff, F., Kalenichenko, D., Philbin, J.: Facenet: A unified embedding for face recognition and clustering. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 815–823 (2015)

    2015

  10. [18]

    ACM Transactions on Database Systems (TODS) 42(3), 1–21 (2017) 18 H

    Schubert, E., Sander, J., Ester, M., Kriegel, H.P., Xu, X.: Dbscan revisited, revisited: why and how you should (still) use dbscan. ACM Transactions on Database Systems (TODS) 42(3), 1–21 (2017) 18 H. Kim et al

    2017

  11. [19]

    Su, Y., Yao, D., Li, J., Wang, B., Bi, J., Zhao, S., Wu, R., Tao, J., Deng, H.: Trajectory-based mobile game bots detection with gaussian mixture model. In: Artificial Neural Networks and Machine Learning–ICANN 2022: 31st International Conference on Artificial Neural Networks,...

    2022

  12. [20]

    In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

    Tao, J., Lin, J., Zhang, S., Zhao, S., Wu, R., Fan, C., Cui, P.: Mvan: Multi-view attention networks for real money trading detection in online games. In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. pp. 2536–2546 (2019)

    2019

  13. [21]

    ACM Transactions on Intelligent Systems and Technology (TIST)12(6), 1–26 (2021)

    Tedjopurnomo, D.A., Li, X., Bao, Z., Cong, G., Choudhury, F., Qin, A.K.: Sim- ilar trajectory search with spatio-temporal deep representation learning. ACM Transactions on Intelligent Systems and Technology (TIST)12(6), 1–26 (2021)

    2021

  14. [22]

    Advances in neural information processing systems 30 (2017)

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł., Polosukhin, I.: Attention is all you need. Advances in neural information processing systems 30 (2017)

    2017

  15. [23]

    In: Proceedings 18th international conference on data engineering

    Vlachos, M., Kollios, G., Gunopulos, D.: Discovering similar multidimensional trajectories. In: Proceedings 18th international conference on data engineering. pp. 673–684. IEEE (2002)

    2002

  16. [24]

    IEEE Transactions on Intelligent Transportation Systems (2024)

    Wang, C., Huang, J., Wang, Y., Lin, Z., Jin, X., Jin, X., Weng, D., Wu, Y.: A deep spatiotemporal trajectory representation learning framework for clustering. IEEE Transactions on Intelligent Transportation Systems (2024)

    2024

  17. [25]

    In: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management

    Wang, H., Li, Z.: Region representation learning via mobility flow. In: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management. pp. 237–246 (2017)

    2017

  18. [26]

    In: International conference on machine learning

    Wang, T., Isola, P.: Understanding contrastive representation learning through alignment and uniformity on the hypersphere. In: International conference on machine learning. pp. 9929–9939. PMLR (2020)

    2020

  19. [27]

    In: Proceedings of the 25th ACM SIGSPATIAL international conference on advances in geographic information systems

    Yan, B., Janowicz, K., Mai, G., Gao, S.: From itdl to place2vec: Reasoning about place type similarity and relatedness by learning embeddings from augmented spatial contexts. In: Proceedings of the 25th ACM SIGSPATIAL international conference on advances in geographic informat...

    2017

  20. [28]

    In: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management

    Yao, D., Zhang, C., Huang, J., Bi, J.: Serm: A recurrent model for next location prediction in semantic trajectories. In: Proceedings of the 2017 ACM on Conference on Information and Knowledge Management. pp. 2411–2414 (2017)

    2017

  21. [29]

    In: Proceedings of the 27th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems

    Yin, Y., Liu, Z., Zhang, Y., Wang, S., Shah, R.R., Zimmermann, R.: Gps2vec: Towards generating worldwide gps embeddings. In: Proceedings of the 27th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. pp. 416–419 (2019)

    2019

  22. [30]

    In: 2022 IEEE 38th International Conference on Data Engineering (ICDE)

    Zhao, S., Fang, J., Zhao, S., Wu, R., Tao, J., Li, S., Pan, G.: T-detector: A trajectory based pre-trained model for game bot detection in mmorpgs. In: 2022 IEEE 38th International Conference on Data Engineering (ICDE). pp. 992–1003. IEEE (2022)

    2022

  23. [31]

    ACM Transactions on Intelligent Systems and Technology (TIST)6(3), 1–41 (2015)

    Zheng, Y.: Trajectory data mining: an overview. ACM Transactions on Intelligent Systems and Technology (TIST)6(3), 1–41 (2015)

    2015

  24. [32]

    IEEE Data Eng

    Zheng, Y., Xie, X., Ma, W.Y., et al.: Geolife: A collaborative social networking service among user, location and trajectory. IEEE Data Eng. Bull.33(2), 32–39 (2010)

    2010

  25. [33]

    In: 2018 IEEE International Conference on Big Data (Big Data)

    Zhou, Y., Huang, Y.: Deepmove: Learning place representations through large scale movement data. In: 2018 IEEE International Conference on Big Data (Big Data). pp. 2403–2412. IEEE (2018)

    2018

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.