Pith. sign in

REVIEW 4 major objections 5 minor 39 references

FairTP: A Prolonged Fairness Framework for Traffic Prediction

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read FairTP aims to make traffic predictions fair across regions and over time by defining sensor states and rebalancing training data.

desk verdict Prolonged fairness is a real gap, but the new SDF metric is self-referential and the evaluation is confounded by uneven sensor budgets; the current evidence doesn't support the paper's headline claims. read the letter →

arxiv 2412.16214 v1 pith:XSX7C2PZ submitted 2024-12-18 cs.LG

Jiangnan Xia , Yu Yang , Jiaxing Shen , Senzhang Wang , Jiannong Cao This is my paper · ORCID
classification cs.LG
keywords trafficpredictionalgorithmicfairnessprolongedregion-basedstaticsensor-baseddynamicstate-guidedbalancedsamplingstateidentificationspatio-temporalgraphneuralnetwork
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

FairTP is a plug-in framework for traffic prediction that aims to make forecasts fair across city regions and over time. Existing fairness methods for traffic typically constrain predictions only at single time points, so fairness can break as traffic conditions change. FairTP defines two fairness metrics: region-based static fairness (RSF), the average absolute difference in prediction-error rates between pairs of regions at a time point, and sensor-based dynamic fairness (SDF), the absolute difference between sensors' accumulated 'sacrifice' and 'benefit' states over a short period. A state identification module labels each sensor as 'sacrifice' or 'benefit' and a state-guided sampling module gives more training weight to sacrificing sensors. Experiments on two real-world traffic datasets with seven base models report that FairTP improves both fairness metrics while keeping mean absolute error roughly unchanged or better.

What carries the argument

The mechanism that carries the argument is the state identification module plus a state-guided balanced sampling loop. A state marker converts each sensor's current MAPE into a binary label by comparison with a threshold recorded from the original model during its own training; a discriminator then learns to predict this label from the spatio-temporal model's hidden representation H_st, so states can be assigned at test time. Every Td batches, the framework accumulates the predicted states into an overall state per sensor, converts the sum into a sampling probability through a sigmoid, combines it with region-level balancing probabilities, and greedily selects the Nsam sensors with the lowest probabilities for the next training period. The fairness objectives are L_RSF, the average absolute regional MAPE difference at each time point, and L_SDF, the average absolute difference of accumulated states over the window, added to the prediction loss as λ1 L_RSF + λ2 L_SDF.

What would settle it

Recompute SDF on the test set using ground-truth state labels (each sensor's actual error compared with the threshold) instead of the discriminator's predictions, and run FairTP with Td = 24 instead of 3; if the reported SDF improvements disappear or reverse in either check, the prolonged-fairness claim is not supported.

Watch

Extended reading notes

Core claim

The paper's central claim is that prolonged fairness in traffic prediction can be defined and optimized without sensitive attributes. Each sensor is placed in one of two states by comparing its mean absolute percentage error (MAPE) with a threshold taken from the unmodified base model: 'benefit' (error below threshold) or 'sacrifice' (error above it). Fairness over a period is achieved when the sums of these states across sensors are similar, which is the sensor-based dynamic fairness (SDF) measure; region-based static fairness (RSF) handles the per-time-point regional gap. The FairTP framework realizes this by co-training a discriminator that infers sensor states from the base model's hidden representations, and by using those states to rebalance training samples every Td batches. On the HK and SD datasets, the paper reports that FairTP attached to DCRNN, AGCRN, GWNET, ASTGCN, DSTAGNN, DGCRN, and D2STGNN lowers RSF and SDF while keeping mean absolute error nearly unchanged or improved in most configurations.

Load-bearing premise

The load-bearing premise is that a sensor's 'sacrifice' or 'benefit' state, defined by comparing its current prediction error rate with a threshold from the original model, can be reliably predicted at test time from the hidden representation, and that summing these predicted states over a three-batch window captures prolonged fairness.

Editorial extensions

If this is right

  • A traffic agency can attach FairTP to an existing spatio-temporal predictor without redesigning the model, because the ST module is replaceable.
  • The fairness metrics need no sensitive attributes such as race or gender, so they apply directly to traffic sensor data.
  • Because sampling is rebalanced every Td batches toward sensors in 'sacrifice', predictions in under-served regions can improve while dense-sensor regions do not lose much accuracy, per the regional performance analysis.
  • The composite loss gives a tunable trade-off between accuracy, static regional fairness, and prolonged sensor fairness via λ1 and λ2.
  • Compared with the static fairness methods FairST and SA-Net on AGCRN and D2STGNN backbones, the paper reports lower regional disparity with comparable or lower MAE.

Reading between the lines

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

  • Going beyond the paper, the threshold used to label 'sacrifice' and 'benefit' is calibrated once on the unmodified model; as FairTP's sampling changes the training distribution, that threshold can become stale, so a natural extension is periodic re-calibration and a check of whether SDF gains persist.
  • Going beyond the paper, the definition of 'prolonged' is tied to the window Td=3 chosen in experiments; testing the framework on daily or weekly windows would show whether the fairness guarantee actually extends to the time scales that matter for transport policy.
  • Going beyond the paper, the same state-identification-plus-rebalancing recipe could be applied to other spatio-temporal forecasting tasks, such as energy or ride-hailing demand, where zones alternate between well-served and under-served error regimes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes FairTP, a framework that augments traffic forecasting models with a region-based static fairness (RSF) loss, a sensor-based dynamic fairness (SDF) loss, a state identification module that labels sensors as "sacrifice" or "benefit" via a discriminator, and a state-guided balanced sampling strategy. The authors claim that on the HK and SD datasets FairTP significantly improves fairness while minimizing accuracy degradation, and they report comparisons across seven traffic prediction backbones, two fairness mitigation baselines, ablations, and parameter analyses.

Significance. The problem of fairness over time in traffic prediction is timely and practically relevant, and the paper makes a concrete attempt by releasing code, using two real-world datasets, and testing seven backbone models. If the claimed effects were reliably established, FairTP would be a useful plug-in for equitable traffic forecasting. However, the current manuscript does not verify the central claim: the SDF metric is computed from the model's own discriminator, no baseline SDF values exist, and the main comparison is confounded by an asymmetry in the amount of training data. As presented, the evidence does not support the conclusion that FairTP improves prolonged fairness without accuracy loss.

major comments (4)
  1. [Comparison With Baselines and Appendix D.3] The comparison is uncontrolled: Appendix D.3 states that "for all baselines without FairTP, we use all sensors for training," while FairTP uses Nsam=200 sampled sensors. Consequently, every FairTP-versus-baseline comparison in Tables 1, 2, and 4 simultaneously changes the amount of training data and the model components, so the reported MAE and RSF differences cannot be attributed to the proposed fairness mechanism. A controlled comparison using the same sensor subset for baselines, or a FairTP variant trained on all sensors, is required before any accuracy or fairness claim can be evaluated.
  2. [Sensor-based Dynamic Fairness and State Identification Module] SDF as defined in Equations (3)-(4) is not an observable property of the predictions; it is computed from the discriminator outputs d_tk_vi, and the state labels are generated by thresholding per-sensor MAPE against the original model's per-round MAPE (Appendix C). Since the same module is optimized by LSDF and used to measure the outcome, the reported SDF improvements in Table 1 and Table 3 are self-referential. The paper explicitly states that baselines "cannot produce SDF outputs," so no baseline SDF values exist to substantiate the prolonged-fairness improvement; an external validation based directly on prediction-error dynamics over time is needed.
  3. [Equation (4) and LSDF] The SDF loss has a degenerate minimizer: if the discriminator outputs a constant state for all sensors, every pairwise difference D_Td[vi] - D_Td[vj] is zero, making LSDF equal to zero, and the sigmoid in Equation (5) then produces uniform sampling probabilities. The paper does not analyze or regularize this collapse. The ablation in Table 3, where SDF swings from 0.06 to 188 when LSDF is removed, only shows that the loss changes discriminator outputs; it does not show that the discriminator's states track actual sacrifice/benefit of sensors.
  4. [Table 1 versus Table 4] The reported accuracy changes are numerically inconsistent. For example, the text discussing Table 1 reports FairTP-DCRNN on HK as having MAE +2.46%, while Table 4 reports -1.20%; FairTP-AGCRN HK MAE is +4.91% in Table 4, whereas the raw MAE values in Table 1 (1.939 vs. 1.957) correspond to approximately -0.9%. These discrepancies undermine the accuracy side of the central claim and must be corrected, with a precise statement of how the percentage changes are computed.
minor comments (5)
  1. [Equation (1)] Equation (1) defines M[yhat_t^rp] as the mean absolute percentage error for region rp at time t, but MAPE cannot be computed from a single time point without specifying an error window or aggregation procedure; please clarify.
  2. [Ablation Study discussion] The text says that when LSDF is removed "the SDF decreases remarkably," but Table 3 shows that removing LSDF (noD) makes SDF much larger (worse); the wording appears to be backwards and should be corrected.
  3. [Table 2] The SDF column in Table 2 contains only dashes and is never discussed; if SDF is not computable for FairST and SA-Net, this should be stated in the caption or text to avoid confusion.
  4. [Equation (9)] The notation in Equation (9) is confusing: d is described as the predicted state and Y as the corresponding label, but the cross-entropy form mixes predicted and label roles; please use consistent notation such as y_pred and y_label.
  5. [Table 3] Table 3 lists "FairTP-G2STGNN," which appears to be a typo for "FairTP-D2STGNN"; please correct it.

Circularity Check

2 steps flagged · score 6.0 of 10

Prolonged-fairness claim rests on a self-referential SDF metric: SDF is computed from FairTP's own state discriminator, and baselines are said to be unable to produce SDF, so no independent SDF improvement is demonstrated.

  1. self definitional [Methodology, Sensor-based Dynamic Fairness, Eq. (3); State identification module; Table 1 comparison paragraph]
    "SDF (vi, vj) =|DTd [vi] − DTd [vj]|, (3) ... where dtk vi is the state of the road sensor vi at time point tk. Its value is given by the state identification module. ... Notably, the calculation of SDF relies on the sensor state predicted by FairTP’s state identification module. Since it is absent in the baselines, they cannot produce SDF outputs."

    The prolonged-fairness metric SDF is defined through the output of FairTP's own discriminator. That discriminator is trained to reproduce labels obtained by thresholding the model's MAPE against a threshold recorded from the unmodified baseline, and LSDF is directly minimized during training. A low reported SDF is therefore not independent evidence of fairness: it is the value of an objective that FairTP actively optimizes, computed with its own internal states. Because the paper states baselines cannot produce SDF, there is no baseline value against which the reported SDF improvements can be measured. The claim of improved prolonged fairness reduces to 'FairTP lowers the SDF value it defines through its own module,' which is partly forced by construction.

  2. fitted input called prediction [Appendix C, Specific Example of State Identification Module]
    "Taking DCRNN as an example, by running DCRNN, the MAPE of each round of training is recorded as the threshold value, and then we apply FairTP to DCRNN and get Fair-DCRNN. In each round of training of Fair-DCRNN, the states of sensors are obtained by comparing MAPE of FairTP-DCRNN to the threshold values."

    The 'state' labels are not ground-truth sensor states but a per-round comparison of the current model's own MAPE against a threshold fitted on the original baseline. Any quantity derived from these labels—the discriminator's predicted states, the sampling probabilities, and hence SDF—is a relabeling of the model's error dynamics relative to a fitted constant, not an independently defined fairness property. Calling this 'state identification' and then reporting SDF as a fairness improvement is a fitted input renamed as a prediction: the reported SDF values measure how well the discriminator reproduces labels that were themselves generated from the baseline-fitted threshold.

full rationale

The RSF and accuracy comparisons are external and independently checkable: RSF is computed from region-level MAPE differences, and MAE/RMSE are standard accuracy metrics, so those parts of the evaluation are not circular. The circularity is concentrated in the paper's novel 'prolonged fairness' claim. SDF, the metric used to support that claim, is defined in Eq. (3) in terms of states 'given by the state identification module.' The state labels are manufactured by thresholding each sensor's current MAPE against a threshold recorded from the unmodified baseline (Appendix C), the discriminator is trained with L_dis to reproduce those labels, and LSDF is minimized as part of FairTP's objective. The paper also explicitly states that baselines cannot produce SDF values, so there is no baseline SDF to improve upon. Thus the reported SDF improvements are not an external validation of prolonged fairness; they are partly an artifact of the framework's own definitions and objective. This is partial circularity, not total: the accuracy and RSF findings stand on independent ground, and no load-bearing self-citation is involved. The score reflects that the central novel claim—prolonged fairness via SDF—is not supported by an independent metric, while the more conventional static-fairness and accuracy comparisons retain empirical content.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The central claim rests on a self-defined notion of sensor state and a normative definition of prolonged fairness. The state labels derive from the model's own accuracy relative to a fitted threshold, and the fairness metric SDF is only computable by the framework itself. The sampling strategy further assumes that state-guided resampling improves fairness without a controlled comparison.

free parameters (4)
  • MAPE threshold for state labels = Per-model and per-round MAPE of the original ST model
    State labels are assigned by comparing each sensor's MAPE to a threshold taken from the unmodified baseline's training rounds; this threshold is dataset- and model-specific and derived from data.
  • Dynamic time length Td = 3
    Selected by parameter analysis from values 2 to 5; controls the sampling frequency and the SDF window.
  • Sampled number Nsam = 200
    Selected by parameter analysis from values 100 to 300; the number of sensors used in FairTP training.
  • Loss weights lambda1 and lambda2 = 0.01 and 0.1
    Set without a systematic sweep; arbitrary hyperparameters that trade off accuracy and fairness terms.
assumptions (4)
  • domain assumption A sensor can be meaningfully classified into binary states 'sacrifice' (low accuracy) and 'benefit' (high accuracy) based on MAPE relative to a threshold.
    Central to the state identification module and SDF; no empirical justification is provided that such a binary classification reflects a natural or fair state.
  • domain assumption Prolonged fairness is achieved when the cumulative states of sensors are similar over a period Td.
    This is a normative definition presented as a given, not derived from any fairness principle; it is the very objective being optimized.
  • domain assumption The hidden representation Hst of the ST model contains sufficient information for the discriminator to infer sensor states without ground truth at test time.
    The discriminator is trained on labels from the current model, and its generalization is assumed without evidence.
  • domain assumption Sampling a subset of sensors (200) with state-guided probabilities can reduce regional performance disparities without harming privileged regions.
    The central mechanism of the proposed framework; relies on the assumption that additional training on sacrifice sensors transfers to better regional accuracy.
invented entities (1)
  • Sensor state (sacrifice/benefit)
    purpose: Labels each sensor as either benefiting from or sacrificing under the current model, used to compute SDF and guide sampling.
    The state is defined via the framework's own threshold and discriminator; no external ground truth or falsifiable prediction outside the paper is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FairTP: A Prolonged Fairness Framework for Traffic Prediction." pith.science (2026). https://pith.science/paper/XSX7C2PZ

@misc{pith2026241216214,
  author       = {Pith},
  title        = {Pith review of: FairTP: A Prolonged Fairness Framework for Traffic Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XSX7C2PZ}},
  note         = {Machine review of arXiv:2412.16214}
}
read the original abstract

Traffic prediction plays a crucial role in intelligent transportation systems. Existing approaches primarily focus on improving overall accuracy, often neglecting a critical issue: whether predictive models lead to biased decisions by transportation authorities. In practice, the uneven deployment of traffic sensors across urban areas results in imbalanced data, causing prediction models to perform poorly in certain regions and leading to unfair decision-making. This imbalance ultimately harms the equity and quality of life for residents. Moreover, current fairness-aware machine learning models only ensure fairness at specific time points, failing to maintain fairness over extended periods. As traffic conditions change, such static fairness approaches become ineffective. To address this gap, we propose FairTP, a framework for prolonged fair traffic prediction. We introduce two new fairness definitions tailored for dynamic traffic scenarios. Fairness in traffic prediction is not static; it varies over time and across regions. Each sensor or urban area can alternate between two states: "sacrifice" (low prediction accuracy) and "benefit" (high prediction accuracy). Prolonged fairness is achieved when the overall states of sensors remain similar over a given period. We define two types of fairness: region-based static fairness and sensor-based dynamic fairness. To implement this, FairTP incorporates a state identification module to classify sensors' states as either "sacrifice" or "benefit," enabling prolonged fairness-aware predictions. Additionally, we introduce a state-guided balanced sampling strategy to further enhance fairness, addressing performance disparities among regions with uneven sensor distributions. Extensive experiments on two real-world datasets demonstrate that FairTP significantly improves prediction fairness while minimizing accuracy degradation.

Figures

Figures reproduced from arXiv: 2412.16214 by the authors.

Figure 1
Figure 1. Illustration of fairness issues in traffic prediction. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The framework of the proposed FairTP. Prolonged Fairness Traffic Prediction Framework In this section, we introduce FairTP, a framework for achiev￾ing prolonged fair traffic prediction as shown in [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Regional performance on two datasets [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (4 more)
Figure 6
Figure 6. Figure 6: Case study. adaptive matrix in the baseline that is able to learn more ST information based on the fixed road sensors. The improve￾ment in RSF is slightly noticeable, likely because the predic￾tive performance of most privileged regions is already quite close, but the …
Figure 4
Figure 4. Figure 4: Effect of dynamic time length Td. (a) FairTP-AGCRN on HK (b) FairTP-AGCRN on SD [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Effect of sampled number Nsam. Ablation Study We further conduct an ablation study to evaluate the contri￾bution of each component in FairTP to the performance gain. We deactivate different components and form the following variants. noS removes the LRSF which does not…
Figure 7
Figure 7. Figure 7: Illustration of datasets partition PeMS platform, comprises 716 road sensors for San Diego County in 2019, and we use sensor readings as traffic fea￾tures. The partitions are shown in [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 29 canonical work pages

  1. [1]

    Bai, L.; Yao, L.; Li, C.; Wang, X.; and Wang, C. 2020. Adaptive graph convolutional recurrent network for traffic forecasting. Advances in neural information processing systems, 33: 17804--17815

    2020

  2. [2]

    Boratto, L.; Fabbri, F.; Fenu, G.; Marras, M.; and Medda, G. 2023. Counterfactual graph augmentation for consumer unfairness mitigation in recommender systems. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, 3753--3757

    work page 2023

  3. [3]

    Caton, S.; and Haas, C. 2024. Fairness in machine learning: A survey. ACM Computing Surveys, 56(7): 1--38

    2024

  4. [4]

    Chai, J.; and Wang, X. 2022. Fairness with adaptive weights. In International Conference on Machine Learning, 2853--2866. PMLR

    work page 2022

  5. [5]

    Choi, J.; Choi, H.; Hwang, J.; and Park, N. 2022. Graph neural controlled differential equations for traffic forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, 6367--6374

    work page 2022

  6. [6]

    Dong, Y.; Ma, J.; Wang, S.; Chen, C.; and Li, J. 2023. Fairness in graph mining: A survey. IEEE Transactions on Knowledge and Data Engineering

    work page 2023

  7. [7]

    Y.; and Chen, H

    Du, X.; Li, Z.; Long, C.; Xing, Y.; Philip, S. Y.; and Chen, H. 2024. FELight: Fairness-Aware Traffic Signal Control Via Sample-Efficient Reinforcement Learning. IEEE Transactions on Knowledge and Data Engineering

    work page 2024

  8. [8]

    Fang, Z.; Long, Q.; Song, G.; and Xie, K. 2021. Spatial-temporal graph ode networks for traffic flow forecasting. In Proceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining, 364--373

    work page 2021

Show all 39 references
  1. [9]

    T.; Saleiro, P.; and Jesus, S

    Ghani, R.; Rodolfa, K. T.; Saleiro, P.; and Jesus, S. 2023. Addressing bias and fairness in machine learning: A practical guide and hands-on tutorial. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 5779--5780

    2023

  2. [10]

    Guo, S.; Lin, Y.; Feng, N.; Song, C.; and Wan, H. 2019. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting. In Proceedings of the AAAI conference on artificial intelligence, volume 33, 922--929

    2019

  3. [11]

    Guo, Z.; Li, J.; Xiao, T.; Ma, Y.; and Wang, S. 2023. Towards fair graph neural networks via graph counterfactual. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, 669--678

    2023

  4. [12]

    Hua, W.; Li, L.; Xu, S.; Chen, L.; and Zhang, Y. 2023. Tutorial on large language models for recommendation. In Proceedings of the 17th ACM Conference on Recommender Systems, 1281--1283

    2023

  5. [13]

    Lan, S.; Ma, Y.; Huang, W.; Wang, W.; Yang, H.; and Li, P. 2022. Dstagnn: Dynamic spatial-temporal aware graph neural network for traffic flow forecasting. In International conference on machine learning, 11906--11917. PMLR

    2022

  6. [14]

    Li, F.; Feng, J.; Yan, H.; Jin, G.; Yang, F.; Sun, F.; Jin, D.; and Li, Y. 2023. Dynamic graph convolutional recurrent network for traffic prediction: Benchmark and solution. ACM Transactions on Knowledge Discovery from Data, 17(1): 1--21

    2023

  7. [15]

    Li, Y.; Yang, Y.; Cao, J.; Liu, S.; Tang, H.; and Xu, G. 2024. Toward Structure Fairness in Dynamic Graph Embedding: A Trend-aware Dual Debiasing Approach. In SIGKDD

    2024

  8. [16]

    Li, Y.; Yu, R.; Shahabi, C.; and Liu, Y. 2018. Diffusion convolutional recurrent neural network: Data-driven traffic forecasting. International Conference on Learning Representations

    2018

  9. [17]

    Mahapatra, D.; Dong, C.; and Momma, M. 2023. Querywise fair learning to rank through multi-objective optimization. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 1653--1664

    2023

  10. [18]

    Mehrabi, N.; Morstatter, F.; Saxena, N.; Lerman, K.; and Galstyan, A. 2021. A survey on bias and fairness in machine learning. ACM computing surveys (CSUR), 54(6): 1--35

    2021

  11. [19]

    Miao, H.; Liu, Z.; Zhao, Y.; Zheng, K.; Zhang, Y.; and Jensen, C. S. 2025. LightTR: A Lightweight Framework for Federated Trajectory Recovery. In ICDE

    2025

  12. [20]

    Miao, H.; Shen, J.; Cao, J.; Xia, J.; and Wang, S. 2022. MBA-STNet: Bayes-enhanced discriminative multi-task learning for flow prediction. IEEE Transactions on Knowledge and Data Engineering

    2022

  13. [21]

    Miao, H.; Zhao, Y.; Guo, C.; Yang, B.; Kai, Z.; Huang, F.; Xie, J.; and Jensen, C. S. 2024. A unified replay-based continuous learning framework for spatio-temporal prediction on streaming data. In ICDE

    2024

  14. [22]

    Shao, Z.; Zhang, Z.; Wei, W.; Wang, F.; Xu, Y.; Cao, X.; and Jensen, C. S. 2022. Decoupled dynamic spatial-temporal graph neural network for traffic forecasting. arXiv preprint arXiv:2206.09112

    2022 arXiv

  15. [23]

    Song, W.; Dong, Y.; Liu, N.; and Li, J. 2022. Guide: Group equality informed individual fairness in graph neural networks. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 1625--1634

    2022

  16. [24]

    Song, Z.; Ma, Y.; and King, I. 2022. Individual Fairness in Dynamic Financial Networks. In NeurIPS 2022 Workshop: New Frontiers in Graph Learning

    2022

  17. [25]

    A.; Bao, Z.; Zheng, B.; Choudhury, F

    Tedjopurnomo, D. A.; Bao, Z.; Zheng, B.; Choudhury, F. M.; and Qin, A. K. 2020. A survey on modern deep neural network for traffic prediction: Trends, methods and challenges. IEEE Transactions on Knowledge and Data Engineering, 34(4): 1544--1561

    2020

  18. [26]

    Wan, M.; Zha, D.; Liu, N.; and Zou, N. 2023. In-processing modeling techniques for machine learning fairness: A survey. ACM Transactions on Knowledge Discovery from Data, 17(3): 1--27

    2023

  19. [27]

    Wang, H.; Xia, J.; Yang, Y.; Wang, S.; and Cao, J. 2024. STS2ANet: Spatio-Temporal Synchronized Sliding Attention Network for Accurate Cross-Day Origin-Destination Prediction. In International Conference on Database Systems for Advanced Applications, 186--202. Springer

    2024

  20. [28]

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

    2019 arXiv

  21. [29]

    Xia, J.; Wang, S.; Wang, X.; Xia, M.; Xie, K.; and Cao, J. 2022. Multi-view Bayesian spatio-temporal graph neural networks for reliable traffic flow prediction. International Journal of Machine Learning and Cybernetics, 1--14

    2022

  22. [30]

    Xia, J.; Wang, S.; Wang, X.; Xia, M.; Xie, K.; and Cao, J. 2024. Multi-view Bayesian spatio-temporal graph neural networks for reliable traffic flow prediction. International Journal of Machine Learning and Cybernetics, 15(1): 65--78

    2024

  23. [31]

    Yan, A.; and Howe, B. 2020. Fairness-aware demand prediction for new mobility. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, 1079--1087

    2020

  24. [32]

    Yang, T.; Xu, Z.; Wang, Z.; and Ai, Q. 2023. FARA: Future-aware Ranking Algorithm for Fairness Optimization. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, 2906--2916

    2023

  25. [33]

    Yang, Y.; Cao, J.; Stojmenovic, M.; Wang, S.; Cheng, Y.; Lum, C.; and Li, Z. 2021. Time-capturing dynamic graph embedding for temporal linkage evolution. IEEE Transactions on Knowledge and Data Engineering, 35(1): 958--971

    2021

  26. [34]

    Yao, H.; Tang, X.; Wei, H.; Zheng, G.; and Li, Z. 2019. Revisiting spatial-temporal similarity: A deep learning framework for traffic prediction. In Proceedings of the AAAI, volume 33, 5668--5675

    2019

  27. [35]

    Ye, J.; Liu, Z.; Du, B.; Sun, L.; Li, W.; Fu, Y.; and Xiong, H. 2022. Learning the evolutionary and multi-scale graph structure for multivariate time series forecasting. In Proceedings of the 28th ACM SIGKDD Conference, 2296--2306

    2022

  28. [36]

    Zheng, Y.; Wang, Q.; Zhuang, D.; Wang, S.; and Zhao, J. 2023 a . Fairness-enhancing deep learning for ride-hailing demand prediction. arXiv preprint arXiv:2303.05698

    2023 arXiv

  29. [37]

    Zheng, Y.; Zhong, L.; Wang, S.; Yang, Y.; Gu, W.; Zhang, J.; and Wang, J. 2023 b . Diffuflow: Robust fine-grained urban flow inference with denoising diffusion model. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, 3505--3513

    2023

  30. [38]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  31. [39]

    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 gl...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.