Pith. sign in

REVIEW 5 major objections 6 minor 85 references

Urban Representation Learning for Fine-grained Economic Mapping: A Semi-supervised Graph-based Approach

T0 review · 5 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read SemiGTX claims that a semi-supervised graph fusing street views, POIs, and mobility flows can map district-level GDP across all three economic sectors, reaching $R^2$ of 0.93–0.96 in the Pearl River Delta.

desk verdict A coherent semi-supervised graph framework for sector-level GDP mapping with genuinely useful data work, but the reported R² numbers are not trustworthy because model selection appears to use test metrics and the masking protocol never rules out test-label leakage. read the letter →

arxiv 2505.11645 v1 pith:AVZOFBGJ submitted 2025-05-16 cs.LG cs.CV

Jinzhou Cao , Xiangxu Wang , Jiashi Chen , Wei Tu , Zhenhui Li , Xindong Yang , Tianhong Zhao , Qingquan Li This is my paper · ORCID
classification cs.LGcs.CV
keywords Semi-supervisedgraphlearningMulti-modalgeospatialdataMulti-taskGDPsectoralmappingFeatureexplainabilityStreetviewimageryUrbanrepresentation
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

Official GDP figures are usually released at coarse administrative scales, hiding where growth and decline occur within a city. The paper claims that a graph-based model, SemiGTX, can map gross domestic product for primary (agriculture), secondary (manufacturing and construction), and tertiary (services) sectors down to district level using street-view images, point-of-interest counts, and mobile phone origin–destination flows, even when most labels are masked during training. The central idea is to pair a small amount of supervised regression with a large amount of spatial self-supervision, so that unlabeled districts learn to agree with their neighboring or containing districts while a few labeled districts anchor the prediction scale. In the Pearl River $\Delta$, reported $R^2$ values reach 0.93, 0.96, and 0.94 for the three sectors, and transfers to Beijing and Chengdu retain high accuracy. If true, this offers a cheaper, faster, and more granular route to economic monitoring in data-scarce settings.

What carries the argument

The load-bearing object is the semi-info loss, a weighted sum $L = \lambda L_{\mathrm{info}} + (1-\lambda)L_{\mathrm{reg}}$ that pairs subgraph-level mutual-information maximization with multi-task regression over unmasked districts. Around it sits a graph encoder: modality-specific pre-encoders (a fine-tuned ViT autoencoder for street views, POI category frequencies, and O-D flows converted into edges), a positional/structural encoder using Laplacian and random-walk encodings, and stacked GraphGPS layers whose message-passing (GAT) and efficient attention (Performer) create both real and virtual edges. The loss factor $\lambda$ lets a single framework interpolate between pure self-supervision and pure supervision, and the ablation evidence shows that neither extreme is optimal.

What would settle it

Run the Pearl River $\Delta$ experiment with the same data but a single held-out split fixed before any hyperparameter tuning, repeated over ten random seeds; if the average tertiary-sector $R^2$ drops to the level of the best GNN baseline, the claimed superiority is contradicted.

Watch

Extended reading notes

Core claim

The paper's claim is that SemiGTX, a semi-supervised graph framework that fuses street-view imagery, POI frequencies, and O-D mobility flows, learns district-level representations accurate enough to map three GDP sectors at once, even when most regression labels are masked. The framework balances a regression loss over unmasked districts with a subgraph-level mutual-information loss, so node representations are pulled into agreement with the districts that contain them while a few supervised districts keep the predictions on the correct economic scale. On 50 Pearl River $\Delta$ districts, the model reports $R^2=0.93$ for primary, $R^2=0.96$ for secondary, and $R^2=0.94$ for tertiary sector value-added, outperforming GNN, transformer, and contrastive baselines. Cross-regional runs in Beijing ($R^2$ of 0.82, 0.89, 0.93) and Chengdu (0.79, 0.80, 0.93) support transferability. Ablations show that the self-supervised term is what rescues the primary sector and that positional and structural encodings each help a different sector.

Load-bearing premise

The reported $R^2$ values are valid because the random 7:1:2 split and the random masking in the regression loss keep every test-set label out of training; if that separation fails, or if test labels influence hyperparameter selection, the numbers no longer measure the method's generalization.

Editorial extensions

If this is right

  • With only a fraction of districts labeled, sector-level GDP maps can be produced at administrative granularity, which matters where official statistics are coarse or delayed.
  • The self-supervised term is not a minor add-on: removing it drops the primary-sector $R^2$ from 0.93 to 0.64 in the reported experiments.
  • Street-view imagery and POIs carry complementary signals: SVI is the stronger cue for primary-sector activity, while POI frequencies help most for secondary and tertiary sectors.
  • Both positional and structural graph encodings are needed; dropping both lowers reported performance by roughly 20%.
  • The framework kept working in Beijing and Chengdu despite different data volumes and economic structures, supporting a claimed transfer across regions.

Reading between the lines

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

  • The same semi-info loss could be applied to other sparse spatial labels, such as poverty rates, unemployment, or land-use classes, where the signal is partly local and partly a function of regional context; the paper does not test this.
  • A natural stress test would be mapping GDP at township or grid level, one administrative tier below the districts used here, since the error analysis suggests long-tailed high-value districts are the hardest to predict.
  • The 7:1:2 split covers only 50 districts, so the method's real sensitivity is to how the label mask is chosen; varying the mask rate and the split would reveal whether the $R^2$ gains persist under extreme label scarcity.
  • Because mobility data are aggregated to weekly averages, incorporating time-of-day or day-type patterns could let the model distinguish residential from commercial districts, a distinction street views alone may miss.
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

5 major / 6 minor

Summary. The paper proposes SemiGTX, a graph-based semi-supervised framework for mapping sectoral GDP at the district level using street view images, POI data, and origin–destination mobility flows. The framework combines modality-specific pre-encoders, positional/structural encodings, GraphGPS layers, a semi-information loss that blends subgraph-level graph infomax with masked regression, and multi-task heads for primary, secondary, and tertiary sectors. Experiments in the Pearl River Delta report R² values of 0.93, 0.96, and 0.94 for the three sectors and claim superiority over GNN, transformer, and contrastive baselines, with additional cross-regional experiments in Beijing and Chengdu. SHAP-based attribution is used to explain modality contributions.

Significance. If the reported results hold under a clean evaluation protocol, SemiGTX would be a practically useful contribution to fine-grained economic mapping, particularly for data-scarce regions: the multi-task sectoral formulation, the semi-supervised loss design, and the explainability analysis are valuable and are backed by extensive ablation studies. The paper also demonstrates serious engineering effort in integrating SVI, POI, and mobility data. However, the central empirical claim is currently not established because the evaluation protocol appears to allow test information to enter both model selection and training, and because the results rest on a single split with no error bars. These issues are load-bearing for the paper's main contribution, which is an empirical performance claim.

major comments (5)
  1. [§4.5.1, Table 4; §4.5.2, Table 5] Hyperparameters are selected on the test set whose metrics are then reported as the headline results. In §4.5.1, the loss factor λ is swept over {0, 0.1, 0.3, 0.5, 0.7, 0.9} and λ = 0.5 is chosen because it gives "the best overall performance" in the reported MAE/MSE/R² table; no validation-based selection is described. In §4.5.2, the P/S encoder configuration is chosen in the same manner. With a 7:1:2 split on only 50 districts, the reported R² values are test-selected maxima rather than unbiased estimates, and the baseline comparison in Table 7 inherits this bias. The authors must perform model selection on the validation split and report repeated-seed or repeated-split statistics, including error bars and significance tests, for the final test evaluation.
  2. [§3.4.3, Eq. (7)] The masking procedure is underspecified. The text says that ground truth values for "certain districts are randomly masked" and defines Lreg over unmasked districts, with N′ described only as the number of unmasked districts. It never states that masking is restricted to training/validation nodes or that test nodes are always masked during training. With random masking over all 50 districts and a 10-district test set, test labels can enter Lreg in some epochs, which would directly invalidate the reported test metrics. The authors must state explicitly that test labels are excluded from Lreg at all times and that masking is applied only within the training split.
  3. [§3.3.3, §3.4.2, Table 7] The transductive graph structure creates an additional leakage channel. The O-D flow edges connect all districts, and GraphGPS performs message passing over real and virtual edges, so label information from training nodes can propagate to test nodes even if test labels are correctly masked from Lreg. The paper does not analyze this channel or provide an inductive or edge-perturbed evaluation to quantify its effect. Since the performance advantage over baselines in Table 7 is the paper's central evidence, this channel must be addressed, for example by reporting results with test nodes disconnected from the graph or by adopting an inductive evaluation protocol.
  4. [§4.5.5, Table 8] The cross-regional experiments in Beijing and Chengdu lack essential protocol details: whether the model is trained from scratch or fine-tuned from the PRD model, the train/validation/test split used in each city, the masking policy, and the hyperparameter selection procedure are not reported. Without these details, the "generality" claim cannot be assessed. The same level of protocol transparency required for the PRD experiments, including error bars over repeated runs, should be provided for both cities.
  5. [§4.3] The ViT pre-encoder training protocol is ambiguous: the paper states "After randomly selecting 40% of the total images for training, the entire dataset is utilized for inference." If the 40% sample is drawn from all districts without regard to the test split, then the pre-encoder has seen street-view images from test districts before the graph model is trained, meaning the features of test nodes are not drawn from a strictly held-out set. Please clarify whether the 40% sample is restricted to training districts or explicitly treat the pre-encoder as a transductive component; if the latter, its contribution to the reported gains should be assessed separately.
minor comments (6)
  1. [§3.4.2, Eq. (4)] There is a typo in Eq. (4): the subscript "MMPN" should be "MPNN" to match the text and the definition of the message-passing component.
  2. [§5.3] The text refers to "the comparison between Beijing and Chengdu in Section 5.3.5," but the correct reference is Section 4.5.5.
  3. [§6, Conclusion] The conclusion contains a duplicated phrase: "SemiGTX's SemiGTX's robust generalization capabilities." Please remove the duplication.
  4. [Table 1] Table 1 uses an unconventional "!" symbol and truncated column headers such as "Signal-" and "Multi-Signal-", which make the comparison matrix difficult to read. Please redesign the table with clear checkmarks and full column names.
  5. [§4.5.3, Table 6] The sentence introducing the modality evaluation says "We then added several standard regressors ... to evaluate the contributions of SVIs and POIs," but the exact procedure for embedding the learned representations and the choice of hyperparameters for Ridge, SVM, RF, and MLP are not stated; please add these implementation details.
  6. [General] The paper does not include a data or code availability statement. Given the sensitivity of the evaluation protocol, a clear statement on data availability and, ideally, open release of code and model configurations is needed for reproducibility.

Circularity Check

2 steps flagged · score 6.0 of 10

Reported R² values are partly selected or fitted: λ is chosen on the reported test table and the random masking in Eq. (7) does not exclude test labels from Lreg.

  1. fitted input called prediction [Section 4.5.1, Table 4; reused as final result in Abstract and Section 6]
    "All data were divided into training, validation, and test sets in a 7:1:2 ratio. ... Among the values tested, λ = 0.5 provides the best overall performance, achieving R² scores of 0.93, 0.96, and 0.94 for the primary, secondary and tertiary sectors, respectively."

    The paper never describes using the validation split to select λ, and the exact R² values shown in Table 4 for λ=0.5 are the same values reported as SemiGTX's final performance in the abstract and conclusions. If Table 4 is computed on the test split, which the text does not rule out and the repeated numbers imply, then λ=0.5 is the argmax of the reported test R² over the grid {0, 0.1, 0.3, 0.5, 0.7, 0.9}. The headline 'prediction' is therefore the maximum of the selection criterion on the target metric, not an independent estimate. The later comparison in Table 7 also fixes λ=0.5 for SemiGTX and its variants, so the claimed superiority is conditional on a hyperparameter that was selected using the same test metrics being reported.

  2. fitted input called prediction [Section 3.4.3, Eq. (7) and masking description]
    "In our transductive semi-supervised setting, ground truth (GT) values for certain districts are randomly masked. The model can only perform regression under the supervision signals from unmasked districts."

    The masking is defined over 'certain districts' with no statement that the 20% test districts are always masked or that the unmasked set is restricted to the training/validation split. Since Lreg is summed over every unmasked district, a randomly unmasked test district contributes its own y_i to the training loss through Eq. (7); its ŷ_i is then fitted to that very label. The reported test R² would then measure training-fit residuals rather than predictions. The transductive graph also allows test labels to propagate through GAT message passing even when masked, compounding the leak. As written, the central empirical claim reduces to fitting unless the omitted exclusion of test labels is supplied.

full rationale

The paper's central contribution is an empirical performance claim: SemiGTX achieves R² of 0.93, 0.96, and 0.94 for sectoral GDP mapping. The derivation chain is not a formal theorem; it is a trained model evaluated on held-out districts. Standard supervised evaluation is not circular, and most of the framework components (GraphGPS, GAT, Performer, GIM, SHAP) are cited from external work, not from the authors' own prior results. The paper's self-citations (Cao et al. 2021, 2024, 2025a, 2025b) appear in motivation and related work and are not load-bearing for the reported accuracy, so no self-citation circularity is found. However, two evaluation-protocol steps make the reported predictions partly reduce to fitting. First, λ=0.5 is selected as 'the best overall performance' in Table 4, and the same R² values are then presented as the model's final result; if Table 4 is test-based, the headline is a test-selected maximum. Second, the random masking in Section 3.4.3 is described without any guarantee that test labels are always masked, so Eq. (7) can include test labels in Lreg, directly fitting the 'predictions' to their ground truth. Both issues are omissions in the text rather than fully documented procedures; if the authors confirm that Table 4 uses a validation split and that test nodes are always masked during training, the circularity would be resolved. As written, the empirical claim is partially forced by construction, warranting a score of 6.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The model introduces no new physical entities. Its load-bearing assumptions are domain assumptions about the predictive content of urban data and two evaluation assumptions about the split and masking protocol. The free parameters λ, α, and the chosen P/S encoder configuration are selected using the test set, which inflates the reported accuracy.

free parameters (3)
  • loss factor λ = 0.5
    Chosen by comparing test-set metrics in Table 4; balances self-supervised and supervised loss terms in Eq. 8.
  • multi-task loss weights α = 0.1, 0.01, 0.001 (primary, secondary, tertiary)
    Set according to relative GDP scales; affects optimization and therefore all reported R² values.
  • P/S encoder configuration = LapPE + RWSE
    Selected based on test-set ablation in Table 5; other configurations yield lower R².
assumptions (6)
  • domain assumption District-level sectoral GDP is predictable from SVI, POI, and O-D flow features aggregated at subdistrict level.
    The entire supervised signal rests on this correlation; introduced in Sections 1 and 3.3.
  • domain assumption The O-D flow graph captures economically meaningful spatial interactions between districts.
    Graph edges are derived from mobile phone flows; if mobility does not track economic ties, graph propagation adds noise. Section 3.3.3.
  • domain assumption ViT embeddings trained by image reconstruction (SSIM 0.91) are sufficient to represent economically relevant visual features.
    The ViT pre-encoder is trained only to reconstruct images, not to predict any economic output. Section 3.3.1, Section 4.3.
  • ad hoc to paper The random 7:1:2 split provides an unbiased test set despite spatial autocorrelation of GDP.
    This is an evaluation assumption, not justified; likely optimistic. Section 4.3.
  • ad hoc to paper The random masking of labels in Section 3.4.3 is confined to the training split so test labels never enter L_reg.
    The paper does not state this restriction, and the claim of test R² requires it.
  • standard math Standard graph and transformer components (GAT, Performer, LapPE, RWSE, DGI) work as described in their source papers.
    Used without modification; Section 3.4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Urban Representation Learning for Fine-grained Economic Mapping: A Semi-supervised Graph-based Approach." pith.science (2026). https://pith.science/paper/AVZOFBGJ

@misc{pith2026250511645,
  author       = {Pith},
  title        = {Pith review of: Urban Representation Learning for Fine-grained Economic Mapping: A Semi-supervised Graph-based Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AVZOFBGJ}},
  note         = {Machine review of arXiv:2505.11645}
}
read the original abstract

Fine-grained economic mapping through urban representation learning has emerged as a crucial tool for evidence-based economic decisions. While existing methods primarily rely on supervised or unsupervised approaches, they often overlook semi-supervised learning in data-scarce scenarios and lack unified multi-task frameworks for comprehensive sectoral economic analysis. To address these gaps, we propose SemiGTX, an explainable semi-supervised graph learning framework for sectoral economic mapping. The framework is designed with dedicated fusion encoding modules for various geospatial data modalities, seamlessly integrating them into a cohesive graph structure. It introduces a semi-information loss function that combines spatial self-supervision with locally masked supervised regression, enabling more informative and effective region representations. Through multi-task learning, SemiGTX concurrently maps GDP across primary, secondary, and tertiary sectors within a unified model. Extensive experiments conducted in the Pearl River Delta region of China demonstrate the model's superior performance compared to existing methods, achieving R2 scores of 0.93, 0.96, and 0.94 for the primary, secondary and tertiary sectors, respectively. Cross-regional experiments in Beijing and Chengdu further illustrate its generality. Systematic analysis reveals how different data modalities influence model predictions, enhancing explainability while providing valuable insights for regional development planning. This representation learning framework advances regional economic monitoring through diverse urban data integration, providing a robust foundation for precise economic forecasting.

Figures

Figures reproduced from arXiv: 2505.11645 by the authors.

Figure 1
Figure 1. Overview of the SemiGTX Framework. Multiple data sources are integrated and [PITH_FULL_IMAGE:figures/full_fig_p014_1.png] view at source ↗
Figure 2
Figure 2. Structure of the fine-tuned ViT pre-encoder for SVIs. Each SVI is processed [PITH_FULL_IMAGE:figures/full_fig_p015_2.png] view at source ↗
Figure 3
Figure 3. The workflow of the Semi-info loss. Positive sample features are transformed into [PITH_FULL_IMAGE:figures/full_fig_p021_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Distribution of the three economic sectors in the mega cities, large cities, and [PITH_FULL_IMAGE:figures/full_fig_p036_4.png]
Figure 5
Figure 5. Figure 5: Geographic distribution of (a-c) value-added and (d-f) model absolute error in [PITH_FULL_IMAGE:figures/full_fig_p037_5.png]
Figure 6
Figure 6. Figure 6: SHAP values on the GDP mapping task. (a) Schematic of layer-wise trace [PITH_FULL_IMAGE:figures/full_fig_p039_6.png]
Figure 7
Figure 7. Figure 7: Importance of input modalities across the three city groups, visualized using POI [PITH_FULL_IMAGE:figures/full_fig_p041_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

85 extracted references · 63 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  3. [3]

    , author Perozzi, B

    author Abu-El-Haija, S. , author Perozzi, B. , author Al-Rfou, R. , author Alemi, A. , year 2018 . title Watch your step: learning node embeddings via graph attention , in: booktitle Proceedings of the 32nd International Conference on Neural Information Processing Systems , publisher Curran Associates Inc. , address Red Hook, NY, USA . p. pages 9198–9208

    2018

  4. [4]

    , author Batra, G

    author Anand, A. , author Batra, G. , year 2022 . title Application of Geospatial Methods in Evaluating Environmental Interventions and Related Socioeconomic Benefits . publisher Springer International Publishing , address Cham . pp. pages 275--289

    2022

  5. [5]

    , author Dertkigil, S.S

    author Bassi, P.R. , author Dertkigil, S.S. , author Cavalli, A. , year 2024 . title Improving deep neural network generalization and robustness to background bias via layer-wise relevance propagation optimization . journal Nature Communications volume 15 , pages 291

    2024

  6. [6]

    , author Niyogi, P

    author Belkin, M. , author Niyogi, P. , year 2001 . title Laplacian eigenmaps and spectral techniques for embedding and clustering , in: booktitle Proceedings of the 15th International Conference on Neural Information Processing Systems: Natural and Synthetic , publisher MIT Press , address Cambridge, MA, USA . p. pages 585–591

    work page 2001

  7. [7]

    , author Carlini, N

    author Berthelot, D. , author Carlini, N. , author Goodfellow, I. , author Oliver, A. , author Papernot, N. , author Raffel, C. , year 2019 . title MixMatch: a holistic approach to semi-supervised learning . publisher Curran Associates Inc. , address Red Hook, NY, USA

    work page 2019

  8. [8]

    , author Song, J

    author Cai, Z. , author Song, J. , author Zhang, T. , author Hu, C. , author Jing, X.Y. , year 2024 . title Local weight coupled network: Multi-modal unequal semi-supervised domain adaptation . journal Multimedia Tools and Applications volume 83 , pages 4331--4357

    work page 2024

Show all 85 references
  1. [9]

    , author Cao, X

    author Cao, J. , author Cao, X. , author Tu, W. , author Tan, X. , author Wang, T. , author Chen, G. , author Zhang, X. , author Li, Q. , year 2025 a. title Nighttime light imagery or mobile phone footprints: Which better reflects urban socio-economics at the grid level? a cas...

    2025

  2. [10]

    , author Li, Q

    author Cao, J. , author Li, Q. , author Tu, W. , author Gao, Q. , author Cao, R. , author Zhong, C. , year 2021 . title Resolving urban mobility networks from individual travel graphs using massive-scale mobile phone tracking data . journal Cities volume 110 , pages 103077 . :...

    2021

  3. [11]

    , author Tu, W

    author Cao, J. , author Tu, W. , author Cao, R. , author Gao, Q. , author Chen, G. , author Li, Q. , year 2024 . title Untangling the association between urban mobility and urban elements . journal Geo-spatial Information Science volume 27 , pages 1071--1089

    2024

  4. [12]

    , author Wang, X

    author Cao, J. , author Wang, X. , author Chen, G. , author Tu, W. , author Shen, X. , author Zhao, T. , author Chen, J. , author Li, Q. , year 2025 b. title Disentangling the hourly dynamics of mixed urban function: A multimodal fusion perspective using dynamic graphs . journ...

    2025

  5. [13]

    , author Tu, W

    author Cao, R. , author Tu, W. , author Cai, J. , author Zhao, T. , author Xiao, J. , author Cao, J. , author Gao, Q. , author Su, H. , year 2022 . title Machine learning-based economic development mapping from multi-source open geospatial data . journal ISPRS Annals of the Ph...

    2022

  6. [14]

    , author Jang, G

    author Chang, B. , author Jang, G. , author Kim, S. , author Kang, J. , year 2020 . title Learning graph-based geographical latent representation for point-of-interest recommendation , in: booktitle Proceedings of the 29th ACM International conference on information & knowledg...

    2020

  7. [15]

    , author Likhosherstov, V

    author Choromanski, K. , author Likhosherstov, V. , author Dohan, D. , author Song, X. , author Gane, A. , author Sarlos, T. , author Hawkins, P. , author Davis, J. , author Mohiuddin, A. , author Kaiser, L. , et al., year 2020 . title Rethinking attention with performers . jo...

    2020 arXiv

  8. [16]

    , author Wang, X

    author Cui, P. , author Wang, X. , author Pei, J. , author Zhu, W. , year 2018 . title A survey on network embedding . journal IEEE transactions on knowledge and data engineering volume 31 , pages 833--852

    2018

  9. [17]

    , author Hadjikakou, M

    author Custodio, H.M. , author Hadjikakou, M. , author Bryan, B.A. , year 2023 . title A review of socioeconomic indicators of sustainability and wellbeing building on the social foundations framework . journal Ecological Economics volume 203 , pages 107608

    2023

  10. [18]

    , author Guan, Y

    author Deng, R. , author Guan, Y. , author Cai, D. , author Yang, T. , author Fraedrich, K. , author Zhang, C. , author Tang, J. , author Liao, Z. , author Wei, Z. , author Guo, S. , year 2023 . title Supervised versus semi-supervised urban functional area prediction: uncertai...

    2023

  11. [19]

    , year 2020

    author Dosovitskiy, A. , year 2020 . title An image is worth 16x16 words: Transformers for image recognition at scale . journal arXiv preprint arXiv:2010.11929

    2020 arXiv

  12. [20]

    , author Joshi, C.K

    author Dwivedi, V.P. , author Joshi, C.K. , author Luu, A.T. , author Laurent, T. , author Bengio, Y. , author Bresson, X. , year 2023 . title Benchmarking graph neural networks . journal Journal of Machine Learning Research volume 24 , pages 1--48

    2023

  13. [21]

    , author Luu, A.T

    author Dwivedi, V.P. , author Luu, A.T. , author Laurent, T. , author Bengio, Y. , author Bresson, X. , year 2021 . title Graph neural networks with learnable structural and positional representations . journal arXiv preprint arXiv:2110.07875

    2021 arXiv

  14. [22]

    , author Sun, H

    author Fan, D. , author Sun, H. , author Yao, J. , author Zhang, K. , author Yan, X. , author Sun, Z. , year 2021 . title Well production forecasting based on arima-lstm model considering manual operations . journal Energy volume 220 , pages 119708

    2021

  15. [23]

    , author Sun, Z

    author Fan, Y. , author Sun, Z. , year 2021 . title Cpi big data prediction based on wavelet twin support vector machine . journal International Journal of Pattern Recognition and Artificial Intelligence volume 35 , pages 2159013

    2021

  16. [24]

    , author Zhang, F

    author Fan, Z. , author Zhang, F. , author Loo, B.P. , author Ratti, C. , year 2023 . title Urban visual intelligence: Uncovering hidden city profiles with street view images . journal Proceedings of the National Academy of Sciences volume 120 , pages e2220417120

    2023

  17. [25]

    , author Leskovec, J

    author Grover, A. , author Leskovec, J. , year 2016 . title node2vec: Scalable feature learning for networks , in: booktitle Proceedings of the 22nd ACM SIGKDD international conference on Knowledge discovery and data mining , pp. pages 855--864

    2016

  18. [26]

    , author Jing, X.Y

    author Jia, X. , author Jing, X.Y. , author Zhu, X. , author Chen, S. , author Du, B. , author Cai, Z. , author He, Z. , author Yue, D. , year 2020 . title Semi-supervised multi-view deep discriminant representation learning . journal IEEE transactions on pattern analysis and ...

    2020

  19. [27]

    , author Liu, L

    author Jin, G. , author Liu, L. , author Li, F. , author Huang, J. , year 2023 . title Spatio-temporal graph neural point process for traffic congestion event prediction , in: booktitle Proceedings of the AAAI Conference on Artificial Intelligence , pp. pages 14268--14276

    2023

  20. [28]

    , author Swanson, N.R

    author Kim, H.H. , author Swanson, N.R. , year 2018 . title Methods for backcasting, nowcasting and forecasting using factor-midas: With an application to korean gdp . journal Journal of Forecasting volume 37 , pages 281--302

    2018

  21. [29]

    , author Mintun, E

    author Kirillov, A. , author Mintun, E. , author Ravi, N. , author Mao, H. , author Rolland, C. , author Gustafson, L. , author Xiao, T. , author Whitehead, S. , author Berg, A.C. , author Lo, W.Y. , et al., year 2023 . title Segment anything , in: booktitle Proceedings of the...

    2023

  22. [30]

    , author Aila, T

    author Laine, S. , author Aila, T. , year 2016 . title Temporal ensembling for semi-supervised learning . journal arXiv preprint arXiv:1610.02242

    2016 arXiv

  23. [31]

    , author Woo, J.O

    author Lee, J. , author Woo, J.O. , author Moon, H. , author Lee, K. , year 2023 . title Unsupervised accuracy estimation of deep visual models using domain-adaptive adversarial perturbation without source samples , in: booktitle Proceedings of the IEEE/CVF International Confe...

    2023

  24. [32]

    , author Xin, S

    author Li, T. , author Xin, S. , author Xi, Y. , author Tarkoma, S. , author Hui, P. , author Li, Y. , year 2022 . title Predicting multi-level socioeconomic indicators from structural urban imagery , in: booktitle Proceedings of the 31st ACM international conference on inform...

    2022

  25. [33]

    , author Huang, W

    author Li, Y. , author Huang, W. , author Cong, G. , author Wang, H. , author Wang, Z. , year 2023 . title Urban region representation learning with openstreetmap building footprints , in: booktitle Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data ...

    2023

  26. [34]

    , author Yu, R

    author Li, Y. , author Yu, R. , author Shahabi, C. , author Liu, Y. , year 2017 . title Diffusion convolutional recurrent neural network: Data-driven traffic forecasting . journal arXiv preprint arXiv:1707.01926

    2017 arXiv

  27. [35]

    , author Biljecki, F

    author Liu, P. , author Biljecki, F. , year 2022 . title A review of spatially-explicit geoai applications in urban geography . journal International Journal of Applied Earth Observation and Geoinformation volume 112 , pages 102936

    2022

  28. [36]

    , author Cheng, D

    author Liu, Y. , author Cheng, D. , author Zhang, D. , author Xu, S. , author Han, J. , year 2025 . title Capsule networks with residual pose routing . journal IEEE Transactions on Neural Networks and Learning Systems volume 36 , pages 2648--2661

    2025

  29. [37]

    , author Li, C

    author Liu, Y. , author Li, C. , author Xu, S. , author Han, J. , year 2024 . title Part-whole relational fusion towards multi-modal scene understanding . http://arxiv.org/abs/2410.14944 arXiv:2410.14944

    2024 arXiv

  30. [38]

    , author Lee, S.I

    author Lundberg, S.M. , author Lee, S.I. , year 2017 . title A unified approach to interpreting model predictions , in: booktitle Proceedings of the 31st International Conference on Neural Information Processing Systems , publisher Curran Associates Inc. , address Red Hook, NY...

    2017

  31. [39]

    , author Chung, F.l

    author Luo, Y. , author Chung, F.l. , author Chen, K. , year 2022 . title Urban region profiling via multi-graph representation learning , in: booktitle Proceedings of the 31st ACM international conference on information & knowledge management , pp. pages 4294--4298

    2022

  32. [40]

    , author Karaku s , O

    author Ma, W. , author Karaku s , O. , author Rosin, P.L. , year 2023 . title Confidence guided semi-supervised learning in land cover classification , in: booktitle IGARSS 2023-2023 IEEE International Geoscience and Remote Sensing Symposium , organization IEEE . pp. pages 5487--5490

    2023

  33. [41]

    , author Specht, A

    author Machicao, J. , author Specht, A. , author Vellenich, D. , author Meneguzzi, L. , author David, R. , author Stall, S. , author Ferraz, K. , author Mabile, L. , author O'Brien, M. , author Corrêa, P. , year 2022 . title A deep-learning method for the prediction of socio-e...

    2022

  34. [42]

    , author Wang, Y

    author Miao, R. , author Wang, Y. , author Li, S. , year 2021 . title Analyzing urban spatial patterns and functional zones using sina weibo poi data: A case study of beijing . journal Sustainability volume 13 , pages 647

    2021

  35. [43]

    , author Mishra, E

    author Murugesan, R. , author Mishra, E. , author Krishnan, A.H. , year 2022 . title Forecasting agricultural commodities prices using deep learning-based models: basic lstm, bi-lstm, stacked lstm, cnn lstm, and convolutional lstm . journal International Journal of Sustainable...

    2022

  36. [44]

    , author Pandey, H

    author Naaz, S. , author Pandey, H. , author Lakshmi, C. , year 2024 . title Forecasting gdp per capita using machine learning algorithms , in: booktitle 2024 Second International Conference on Emerging Trends in Information Technology and Engineering (ICETITE) , organization ...

    2024

  37. [45]

    , author Pedreschi, D

    author Pappalardo, L. , author Pedreschi, D. , author Smoreda, Z. , author Giannotti, F. , year 2015 . title Using big data to study the link between human mobility and socio-economic development , in: booktitle 2015 IEEE international conference on big data (big data) , organ...

    2015

  38. [46]

    , author Al-Rfou, R

    author Perozzi, B. , author Al-Rfou, R. , author Skiena, S. , year 2014 . title Deepwalk: Online learning of social representations , in: booktitle Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining , pp. pages 701--710

    2014

  39. [47]

    , author Atkinson, P.M

    author Price, N. , author Atkinson, P.M. , year 2022 . title Global gdp prediction with night-lights and transfer learning . journal IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing volume 15 , pages 7128--7138

    2022

  40. [48]

    , author Fawcett, T

    author Provost, F. , author Fawcett, T. , year 2013 . title Data science and its relationship to big data and data-driven decision making . journal Big data volume 1 , pages 51--59

    2013

  41. [49]

    , author Martinez Jr, A

    author Puttanapong, N. , author Martinez Jr, A. , author Bulan, J.A.N. , author Addawe, M. , author Durante, R.L. , author Martillan, M. , year 2022 . title Predicting poverty using geospatial data in thailand . journal ISPRS International Journal of Geo-Information volume 11 ...

    2022

  42. [50]

    , author Xiao, M

    author Qiao, Z. , author Xiao, M. , author Guo, W. , author Luo, X. , author Xiong, H. , year 2024 . title Information filtering and interpolating for semi-supervised graph domain adaptation . journal Pattern Recognition volume 153 , pages 110498

    2024

  43. [51]

    , author van Florenstein Mulder , T

    author Richardson, A. , author van Florenstein Mulder , T. , author Vehbi, T. , year 2021 . title Nowcasting gdp using machine-learning algorithms: A real-time assessment . journal International Journal of Forecasting volume 37 , pages 941--948

    2021

  44. [52]

    , author Saul, L.K

    author Roweis, S.T. , author Saul, L.K. , year 2000 . title Nonlinear dimensionality reduction by locally linear embedding . journal science volume 290 , pages 2323--2326

    2000

  45. [53]

    , author Cogswell, M

    author Selvaraju, R.R. , author Cogswell, M. , author Das, A. , author Vedantam, R. , author Parikh, D. , author Batra, D. , year 2017 . title Grad-cam: Visual explanations from deep networks via gradient-based localization , in: booktitle 2017 IEEE International Conference on...

    2017

  46. [54]

    , author Wang, Y

    author Shi, B. , author Wang, Y. , author Guo, F. , author Xu, B. , author Shen, H. , author Cheng, X. , year 2024 . title Graph domain adaptation: Challenges, progress and prospects . journal arXiv preprint arXiv:2402.00904

    2024 arXiv

  47. [55]

    , author Wu, Y

    author Shi, K. , author Wu, Y. , author Li, D. , author Li, X. , year 2022 . title Population, gdp, and carbon emissions as revealed by snpp-viirs nighttime light data in china with different scales . journal IEEE Geoscience and Remote Sensing Letters volume 19 , pages 1--5

    2022

  48. [56]

    , author Namin, A.S

    author Siami-Namini, S. , author Namin, A.S. , year 2018 . title Forecasting economics and financial time series: Arima vs. lstm . journal arXiv preprint arXiv:1803.06386

    2018 arXiv

  49. [57]

    , author Watson, M.W

    author Stock, J.H. , author Watson, M.W. , year 2001 . title Vector autoregressions . journal Journal of Economic perspectives volume 15 , pages 101--115

    2001

  50. [58]

    , author Shi, Z

    author Sun, Y. , author Shi, Z. , author Li, Y. , year 2023 . title A graph-theoretic framework for understanding open-world semi-supervised learning , in: booktitle Proceedings of the 37th International Conference on Neural Information Processing Systems , publisher Curran As...

    2023

  51. [59]

    , author Qu, M

    author Tang, J. , author Qu, M. , author Wang, M. , author Zhang, M. , author Yan, J. , author Mei, Q. , year 2015 . title Line: Large-scale information network embedding , in: booktitle Proceedings of the 24th international conference on world wide web , pp. pages 1067--1077

    2015

  52. [60]

    U lker, E.D. , author \

    author \"U lker, E.D. , author \"U lker, S. , year 2019 . title Unemployment rate and gdp prediction using support vector regression , in: booktitle Proceedings of the 1st International Conference on Advanced Information Science and System , pp. pages 1--5

    2019

  53. [61]

    , author Cucurull, G

    author Velickovic, P. , author Cucurull, G. , author Casanova, A. , author Romero, A. , author Lio, P. , author Bengio, Y. , et al., year 2017 . title Graph attention networks . journal stat volume 1050 , pages 10--48550

    2017

  54. [62]

    , author Fedus, W

    author Veli c kovi \'c , P. , author Fedus, W. , author Hamilton, W.L. , author Li \`o , P. , author Bengio, Y. , author Hjelm, R.D. , year 2018 . title Deep graph infomax . journal arXiv preprint arXiv:1809.10341

    2018 arXiv

  55. [63]

    , author Essi, I.D

    author Victor-Edema, U.A. , author Essi, I.D. , year 2016 . title Autoregressive integrated moving average with exogenous variable (arimax) model for nigerian non-oil export . journal European Journal of Business and Management volume 8 , pages 29--34

    2016

  56. [64]

    , author Wang, X

    author Wang, D. , author Wang, X. , author Xiang, Z. , author Yu, D. , author Deng, S. , author Xu, G. , year 2021 . title Attentive sequential model based on graph neural network for next poi recommendation . journal World Wide Web volume 24 , pages 2161--2184

    2021

  57. [65]

    , author Li, Z

    author Wang, H. , author Li, Z. , year 2017 . title Region representation learning via mobility flow , in: booktitle Proceedings of the 2017 ACM on Conference on Information and Knowledge Management , pp. pages 237--246

    2017

  58. [66]

    , author Yang, S

    author Wang, W. , author Yang, S. , author He, Z. , author Wang, M. , author Zhang, J. , author Zhang, W. , year 2018 . title Urban perception of commercial activeness from satellite images and streetscapes , in: booktitle Companion Proceedings of the The Web Conference 2018 ,...

    2018

  59. [67]

    , author Jing, C

    author Wang, Y. , author Jing, C. , author Huang, W. , author Jin, S. , author Lv, X. , year 2023 . title Adaptive spatiotemporal inceptionnet for traffic flow forecasting . journal IEEE Transactions on Intelligent Transportation Systems volume 24 , pages 3882--3907

    2023

  60. [68]

    , author Zhu, D

    author Wang, Y. , author Zhu, D. , year 2024 . title A hypergraph-based hybrid graph convolutional network for intracity human activity intensity prediction and geographic relationship interpretation . journal Information Fusion volume 104 , pages 102149

    2024

  61. [69]

    , author Bovik, A.C

    author Wang, Z. , author Bovik, A.C. , author Sheikh, H.R. , author Simoncelli, E.P. , year 2004 . title Image quality assessment: from error visibility to structural similarity . journal IEEE transactions on image processing volume 13 , pages 600--612

    2004

  62. [70]

    , author Zheng, J

    author Wang, Z. , author Zheng, J. , author Han, C. , author Lu, B. , author Yu, D. , author Yang, J. , author Han, L. , year 2025 . title A comprehensive assessment approach for multiscale regional economic development: Fusion modeling of nighttime lights and openstreetmap da...

    2025

  63. [71]

    , author Meyer, H

    author Wei, D. , author Meyer, H. , author Bacchin, T.K. , year 2023 . title Spatial dynamics in the pearl river delta and development strategies , in: booktitle Adaptive Urban Transformation: Urban Landscape Dynamics, Regional Design and Territorial Governance in the Pearl Ri...

    2023

  64. [72]

    , author Yan, X

    author Wu, S. , author Yan, X. , author Fan, X. , author Pan, S. , author Zhu, S. , author Zheng, C. , author Cheng, M. , author Wang, C. , year 2022 . title Multi-graph fusion networks for urban region embedding . journal arXiv preprint arXiv:2201.09760

    2022 arXiv

  65. [73]

    , author Zhang, Z

    author Wu, X. , author Zhang, Z. , author Chang, H. , author Huang, Q. , year 2021 . title A data-driven gross domestic product forecasting model based on multi-indicator assessment . journal Ieee Access volume 9 , pages 99495--99503

    2021

  66. [74]

    , author Pan, S

    author Wu, Z. , author Pan, S. , author Long, G. , author Jiang, J. , author Zhang, C. , year 2019 . title Graph wavenet for deep spatial-temporal graph modeling . journal arXiv preprint arXiv:1906.00121

    2019 arXiv

  67. [75]

    , author Zhang, T

    author Xu, C. , author Zhang, T. , author Zhang, D. , author Zhang, D. , author Han, J. , year 2024 . title Deep generative adversarial reinforcement learning for semi-supervised segmentation of low-contrast and small objects in medical images . journal IEEE Transactions on Me...

    2024

  68. [76]

    , author Zhou, B

    author Xu, Y. , author Zhou, B. , author Jin, S. , author Xie, X. , author Chen, Z. , author Hu, S. , author He, N. , year 2022 . title A framework for urban land use classification by integrating the spatial context of points of interest and graph convolutional neural network...

    2022

  69. [77]

    , author Ai, T

    author Yan, X. , author Ai, T. , author Yang, M. , author Yin, H. , year 2019 . title A graph convolutional neural network for classification of building patterns using spatial vector data . journal ISPRS journal of photogrammetry and remote sensing volume 150 , pages 259--273

    2019

  70. [78]

    , author Song, Z

    author Yang, X. , author Song, Z. , author King, I. , author Xu, Z. , year 2022 . title A survey on deep semi-supervised learning . journal IEEE transactions on knowledge and data engineering volume 35 , pages 8934--8954

    2022

  71. [79]

    , author Zhu, Q

    author Yao, Y. , author Zhu, Q. , author Guo, Z. , author Huang, W. , author Zhang, Y. , author Yan, X. , author Dong, A. , author Jiang, Z. , author Liu, H. , author Guan, Q. , year 2023 . title Unsupervised land-use change detection using multi-temporal poi embedding . journ...

    2023

  72. [80]

    , author Fu, Y

    author Yao, Z. , author Fu, Y. , author Liu, B. , author Hu, W. , author Xiong, H. , year 2018 . title Representing urban functions through zone embedding with human mobility patterns , in: booktitle Proceedings of the 27th International Joint Conference on Artificial Intellig...

    2018

  73. [81]

    , author Zhou, X

    author Yong, X. , author Zhou, X. , year 2024 . title Musecl: Predicting urban socioeconomic indicators via multi-semantic contrastive learning . journal arXiv preprint arXiv:2407.09523

    2024

  74. [82]

    , year 2021

    author Yoon, J. , year 2021 . title Forecasting of real gdp growth using machine learning models: Gradient boosting and random forest approach . journal Computational Economics volume 57 , pages 247--265

    2021

  75. [83]

    , author Yin, H

    author Yu, B. , author Yin, H. , author Zhu, Z. , year 2017 . title Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting . journal arXiv preprint arXiv:1709.04875

    2017 arXiv

  76. [84]

    , author Li, T

    author Zhang, M. , author Li, T. , author Li, Y. , author Hui, P. , year 2021 . title Multi-view joint graph representation learning for urban region embedding , in: booktitle Proceedings of the twenty-ninth international conference on international joint conferences on artifi...

    2021

  77. [85]

    , author Du, S

    author Zhang, X. , author Du, S. , author Wang, Q. , year 2017 . title Hierarchical semantic cognition for urban functional zones with vhr satellite images and poi data . journal ISPRS Journal of Photogrammetry and Remote Sensing volume 132 , pages 170--184

    2017

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.