First-author work on self-supervised trajectory, path and urban region representation learning at Aalborg University
3
papers here
3
reviewed by Evolve
3
first-author
0
corresponding
—
works · OpenAlex
—
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-16) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
Output data is not yet available for scoring. This axis would measure total publication volume — papers per year and across venues — which cannot be inferred from the three catalogued 2026 preprints alone.
Impact—
Impact cannot yet be assessed: no citation counts, h-index, or venue-level influence data are provided. This axis would measure citations, citing literature, and downstream adoption of the proposed methods.
Focus0/100
The focus score registers 0/100 because the catalogued output splits evenly across three subfield labels — machine learning for intelligent transportation systems, machine learning / urban computing, and spatio-temporal data mining — each holding about one third (0.33) of the field distribution. In practice the three clusters are thematically adjacent, all centred on representation learning for movement trajectories and urban data.
Collaboration50/100
The collaboration score is 50/100, based on 11 distinct co-authors in the snapshot. The frequent-coauthor list is anchored by Aalborg colleagues — Christian S. Jensen, Hua Lu, Bin Yang, Kristian Torp and Torben Bach Pedersen — alongside Zongyi Xu, Jilin Hu, Ying Sun, Hao Miao, Xiangmeng Wang and Tung Kieu; all three catalogued papers are first-authored.
Activity—
Activity data is insufficient for scoring. This axis would measure publishing cadence and recency — outputs per year and gaps between them — which the snapshot's three 2026 papers alone cannot establish.
Reproducibility70/100
Reproducibility scores 70/100, an average computed over 3 reviewed papers. Evolve rates DGCPath as high readiness, with CURE and REFINE at medium. The work favors broad empirical comparisons — 10 baselines on 3 trajectory datasets for DGCPath, 4 datasets for REFINE and 3 cities for CURE — yet reviews consistently flag missing statistical significance testing.
Research summary
Sean Bin Yang is a researcher at Aalborg University working on representation learning for spatio-temporal data, spanning trajectory, path, and urban region representation learning. Evolve's catalog lists three 2026 first-authored preprints. REFINE unifies contrastive and generative trajectory representation learning through closed-loop transcription, with broad wins across four datasets. DGCPath couples diffusion-generated views with a Jensen–Shannon divergence-based distributional contrastive objective, reported to outperform ten baselines on three trajectory datasets. CURE builds confounder-aware multi-view urban region embeddings via heuristic shared-component residualization, with strong gains on three cities. Across the three papers, Evolve's reviews credit broad empirical coverage but repeatedly flag thin novelty or conditional theory and missing statistical significance tests; reproducibility readiness is rated high for DGCPath and medium for CURE and REFINE, averaging 70/100. The snapshot records 11 distinct co-authors anchored at Aalborg, and an even three-way field split across intelligent transportation systems, urban computing, and spatio-temporal data mining. Output, impact, and activity metrics are not yet available.
Collaboration pattern
Collaboration is mid-range (50/100) with 11 distinct co-authors recorded. The frequent-coauthor list is anchored at Aalborg University — Christian S. Jensen, Hua Lu, Bin Yang, Kristian Torp and Torben Bach Pedersen — alongside Zongyi Xu, Jilin Hu, Ying Sun, Hao Miao, Xiangmeng Wang and Tung Kieu. All three catalogued papers are first-authored (non-corresponding), suggesting a lead-author role within a stable spatio-temporal data group.
Research keywords
Highlights
Confounder-aware multi-view urban region embeddings (CURE)
DGCPath: distribution-aware contrastive path representations
REFINE: closed-loop transcription for trajectory representation learning
Timeline
- 2026First-authored preprint CURE: confounder-aware multi-view urban region representation learning via shared-component residualization.
- 2026First-authored preprint DGCPath: diffusion-generated views with Jensen–Shannon distributional contrastive learning for self-supervised path representations.
- 2026First-authored extended paper REFINE: closed-loop transcription unifying contrastive and generative trajectory representation learning (identified by Evolve as KDD'26).
Papers in the Evolve catalog
- first author
When Correlations Mislead: Confounder-Aware Multi-View Urban Region Representation Learning
- first author
DGCPath: Distribution-Aware Generative Contrastive Framework for Self-supervised Path Representation Learning -- Extended Version
- first author
REFINE: Trajectory Representation Learning via Closed-Loop Transcription -- Extended Version
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometric scores are computed from a small snapshot; the output, impact, and activity dimensions lack underlying productivity and citation data and are therefore not scored.
- The catalog covers only three arXiv papers reviewed by Evolve; the researcher's full publication record may be broader.
- Country metadata is missing, and venue status is only partly indicated (REFINE is described in review as a KDD'26 extended paper).
- Profile is name-resolved from public bibliographic data linked to ORCID 0000-0001-7819-2290; name variants may exist in indexing.