pith. sign in

arxiv: 2605.16447 · v2 · pith:CXJXHNI2new · submitted 2026-05-15 · 💻 cs.LG · cs.AI

Nested Spatio-Temporal Time Series Forecasting

Yinghao Ai , Yukai Zhou , Ruoxi Jiang , Junyi An , Chao Qu , Zhijian Zhou , Shiyu Wang , Fenglei Cao
show 3 more authors
Zenglin Xu Furao Shen Yuan Qi
This is my paper

Pith reviewed 2026-05-20 20:32 UTC · model grok-4.3

classification 💻 cs.LG cs.AI
keywords spatiotemporal forecastingnested frameworkspectral clusteringtime series predictionmacro-micro couplingcoarse-to-fine predictionregional trends
0
0 comments X

The pith

Nested framework couples future macro regional trends with micro observations to guide fine-grained spatio-temporal forecasts

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

This paper establishes a nested forecasting framework for spatio-temporal time series that links predicted large-scale regional trends with detailed local historical observations. It uses spectral clustering to form coherent regions that filter noise while retaining key patterns, then applies a progressive coarse-to-fine predictor to incorporate those future trends during inference. This setup lets the model anticipate changes like periodic offsets rather than relying solely on past spatial priors that often fail under noisy or shifting conditions. Readers working on applications such as traffic management would care because the method claims more reliable fine-grained predictions on high-dimensional data.

Core claim

The nested forecasting framework couples future macro-level regional trends with micro-level historical observations, enabling top-down guidance from abstract future representations for fine-grained forecasting. Spectral clustering constructs semantically coherent regions that filter systematic noise while preserving essential trends, and a progressive coarse-to-fine predictor integrates these representative features to anticipate dynamic anomalies such as periodic offsets.

What carries the argument

Nested forecasting framework that integrates macro future trends with micro observations via spectral clustering for region construction and progressive coarse-to-fine prediction

If this is right

  • The model anticipates dynamic anomalies such as periodic offsets before they appear in the fine-grained data.
  • Performance consistently exceeds state-of-the-art baselines across multiple high-dimensional datasets.
  • Top-down guidance from predicted future representations improves accuracy on tasks that suffer from evolving temporal correlations.

Where Pith is reading between the lines

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

  • The same macro-to-micro nesting idea could be tested on non-spatial hierarchical time series such as multi-scale sales or energy demand data.
  • Replacing spectral clustering with a learned grouping module might allow the framework to adapt regions during training.
  • The approach highlights a general strategy of using coarse future predictions to correct biases in historical features for other forecasting domains.

Load-bearing premise

Spectral clustering on the input data produces regions that are semantically coherent enough to filter systematic noise without discarding essential underlying trends.

What would settle it

Running the method on a high-dimensional spatio-temporal dataset where spectral clustering yields incoherent or noisy regions and measuring whether accuracy still exceeds standard baselines.

Figures

Figures reproduced from arXiv: 2605.16447 by Chao Qu, Fenglei Cao, Furao Shen, Junyi An, Ruoxi Jiang, Shiyu Wang, Yinghao Ai, Yuan Qi, Yukai Zhou, Zenglin Xu, Zhijian Zhou.

Figure 1
Figure 1. Figure 1: Diagram of NEST. (i) Spectral representation: Node-level time series are partitioned via spectral clustering to derive representative regional dynamics Z, which serve as macroscopic structural anchors for the system. (ii) Training phase: We process historical node signals Xt−L+1:t alongside future regional guidance Zt+1:t+P using decoupled encoders. To bridge the discrepancy between training and inference,… view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the Nested Spatio-Temporal Forecasting mechanism. The left panel shows the spatial distribution of two learned latent regions. The right panels visualize the two-stage decoding process: (1) Macro Generation, where the model first predicts a stable regional trend (see top row); and (2) Micro Guidance, where this trend explicitly guides the forecasting of downstream nodes (bottom rows, indica… view at source ↗
Figure 3
Figure 3. Figure 3: Forecasting performance with varying numbers of latent regions (M) on GBA and GLA. 6. Conclusion In this paper, we proposed NEST, a novel macro-to-micro framework designed to tackle local noise sensitivity and cascading error accumulation in high-dimensional spatio￾temporal forecasting. By leveraging data-driven semantic clustering without relying on physical priors, NEST ex￾tracts stable regional represen… view at source ↗
Figure 4
Figure 4. Figure 4: Quantile forecasting under different noise regimes. We visualize the quantile prediction results using three quantile levels {0.1, 0.5, 0.9}, where the shaded region denotes the predictive distribution spanned by the lower and upper quantiles. In the high-noise case (left), the quantile band widens to reflect increased uncertainty and covers most of the ground-truth trajectory, indicating well-calibrated u… view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative comparison of multi-step forecasting results on GBA dataset. We visualize the predictions for two distinct nodes: Node 3195 (Left) and Node 2264 (Right). (a) Top Panels: The macro-level regional trends predicted by our auxiliary module, which serve as guidance. (b) Bottom Panels: The comparison of fine-grained node predictions. The baseline PATCHSTG (blue dashed line) struggles to capture sharp… view at source ↗
Figure 6
Figure 6. Figure 6: Visual comparison between raw node sequences (exhibiting high-frequency noise) and their corresponding macro centroid features (preserving smooth trends). 20 [PITH_FULL_IMAGE:figures/full_fig_p020_6.png] view at source ↗
read the original abstract

Spatiotemporal forecasting is critical for real-world applications like traffic management, yet capturing reliable interactions remains challenging under noisy and non-stationary conditions. Existing methods primarily rely on historical spatial priors, often failing to account for evolving temporal correlations and suffering from systematic errors. In this work, we propose a nested forecasting framework that couples future macro-level regional trends with micro-level historical observations, enabling top-down guidance from abstract future representations for fine-grained forecasting. Specifically, we employ a spectral clustering-based approach to construct semantically coherent regions, providing both theoretical and empirical evidence that this representation effectively filters systematic noise while preserving essential trends. Building on this, we develop a progressive coarse-to-fine predictor to integrate these representative features into the inference process. This enables the model to leverage trend predictions to anticipate dynamic anomalies, such as periodic offsets, in advance. Furthermore, extensive experiments on multiple high-dimensional datasets demonstrate that our method consistently outperforms state-of-the-art baselines, validating the effectiveness of future macro-guided nested forecasting.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 1 minor

Summary. The manuscript presents a nested spatio-temporal time series forecasting framework. It employs spectral clustering on historical data to construct macro-level regions that provide semantically coherent representations, then uses a progressive coarse-to-fine predictor to couple predicted future macro trends with micro-level observations for fine-grained forecasting. The work claims this filters systematic noise while preserving trends, enables anticipation of dynamic anomalies, and consistently outperforms state-of-the-art baselines on multiple high-dimensional datasets, supported by theoretical and empirical evidence for the clustering step.

Significance. If the results hold under rigorous validation, the nested framework could advance handling of noisy and non-stationary spatio-temporal data by providing top-down future guidance, with potential utility in domains such as traffic management. The explicit provision of both theoretical and empirical support for the noise-filtering property of the region construction is a methodological strength worth highlighting.

major comments (2)
  1. Abstract and experimental evaluation: the abstract asserts consistent outperformance and noise-filtering benefits but provides no quantitative details, error bars, or ablation results; the central claims rest on unspecified experiments and post-hoc region construction whose impact on final metrics is not shown.
  2. Section on spectral clustering for region construction: the framework depends on clusters derived from past observations to generate future macro trends that guide micro forecasts. However, under non-stationarity, evolving temporal correlations may render historical clusters non-predictive, causing the coarse-to-fine predictor to propagate misaligned signals rather than filter noise; the manuscript must demonstrate cluster stability across time shifts via temporal hold-out validation or adaptive mechanisms.
minor comments (1)
  1. Ensure all notation for macro- and micro-level features is defined consistently in the methods section and used uniformly in equations and figures.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive feedback. We address each major comment point by point below, indicating where revisions have been made to strengthen the manuscript.

read point-by-point responses

  1. Referee: Abstract and experimental evaluation: the abstract asserts consistent outperformance and noise-filtering benefits but provides no quantitative details, error bars, or ablation results; the central claims rest on unspecified experiments and post-hoc region construction whose impact on final metrics is not shown.

    Authors: We agree that the abstract would benefit from greater specificity to support the central claims. The full manuscript contains quantitative results across multiple datasets, including performance tables with standard deviations, ablation studies isolating the contribution of the nested components, and analysis of the region construction step. To address the concern directly, we have revised the abstract to include key quantitative highlights (e.g., average relative improvements and mention of ablation outcomes) while preserving its concise nature. We have also expanded the experimental section with an explicit subsection quantifying the impact of the spectral region construction on final forecasting metrics. revision: yes

  2. Referee: Section on spectral clustering for region construction: the framework depends on clusters derived from past observations to generate future macro trends that guide micro forecasts. However, under non-stationarity, evolving temporal correlations may render historical clusters non-predictive, causing the coarse-to-fine predictor to propagate misaligned signals rather than filter noise; the manuscript must demonstrate cluster stability across time shifts via temporal hold-out validation or adaptive mechanisms.

    Authors: This is a valid concern regarding the robustness of fixed historical clusters under non-stationarity. Our manuscript already supplies theoretical analysis demonstrating that the spectral clustering filters systematic noise while preserving trends, together with empirical support on the evaluated datasets. To directly validate stability, we have added temporal hold-out experiments in the revised manuscript: clusters are constructed on earlier time windows and assessed for consistency and predictive utility on subsequent held-out periods. These new results confirm that the regions remain stable and continue to provide effective top-down guidance, and we have incorporated a brief discussion of this validation in the clustering section. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The abstract and provided description outline a nested forecasting framework that first applies spectral clustering to historical data for region construction, then uses a progressive coarse-to-fine predictor to integrate macro trends with micro observations. No equations or steps are shown that define a quantity in terms of itself, rename a fitted parameter as a prediction, or reduce the central claim to a self-citation chain. The paper states it supplies both theoretical and empirical evidence for the clustering step and validates performance on external high-dimensional datasets, keeping the derivation self-contained against independent benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the effectiveness of spectral clustering for noise filtering and the benefit of future macro guidance; these are presented as domain assumptions without independent external benchmarks in the abstract.

axioms (1)
  • domain assumption Spectral clustering produces semantically coherent regions that filter systematic noise while preserving essential trends.
    Invoked in abstract as providing both theoretical and empirical evidence for the representation.

pith-pipeline@v0.9.0 · 5725 in / 1262 out tokens · 54461 ms · 2026-05-20T20:32:15.106034+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

103 extracted references · 103 canonical work pages

  1. [1]

    A Survey on Spatio-Temporal Prediction: From Transformers to Foundation Models , volume =

    Mao, Yingchi and Zhou, Hongliang and Chen, Ling and Qi, Rongzhi and Sun, Zhende and Rong, Yi and He, Xiaoming and Chen, Mingkai and Mumtaz, Shahid and Frascolla, Valerio and Guizani, Mohsen and Rodrigues, Joel , journal =. A Survey on Spatio-Temporal Prediction: From Transformers to Foundation Models , volume =

    work page

  2. [2]

    , journal =

    Bassett, Koenker Gilbert and Jr. , journal =. Regression Quantiles , volume =

    work page

  3. [3]

    Author, N. N. , title =

    work page

  4. [4]

    Hierarchical Spatial-Temporal Neural Network with Attention Mechanism for Traffic Flow Forecasting , volume =

    Lian, Qingyun and Sun, Wei and Dong, Wei , journal =. Hierarchical Spatial-Temporal Neural Network with Attention Mechanism for Traffic Flow Forecasting , volume =

    work page

  5. [5]

    Youngseog Chung and Willie Neiswanger and Ian Char and Jeff Schneider , title =

    work page

  6. [6]

    R. O. Duda and P. E. Hart and D. G. Stork , edition =. Pattern Classification , year =

    work page

  7. [7]

    M. J. Kearns , school =. Computational Complexity of Machine Learning , year =

    work page

  8. [8]

    Langley , booktitle =

    P. Langley , booktitle =. Crafting Papers on Machine Learning , year =

    work page

  9. [9]

    HSFE: A hierarchical spatial-temporal feature enhanced framework for traffic flow forecasting , volume =

    Jungang Lou and Xinye Zhang and Ruiqin Wang and Zhenfang Liu and Kang Zhao and Qing Shen , journal =. HSFE: A hierarchical spatial-temporal feature enhanced framework for traffic flow forecasting , volume =

    work page

  10. [10]

    Olivares and Boris N

    Cristian Challu and Kin G. Olivares and Boris N. Oreshkin and Federico Garza Ram

    work page

  11. [11]

    Ilya Loshchilov and Frank Hutter , title =

    work page

  12. [12]

    Huiqiang Wang and Jian Peng and Feihu Huang and Jince Wang and Junhui Chen and Yifei Xiao , title =

    work page

  13. [13]

    Ivan Marisca and Cesare Alippi and Filippo Maria Bianchi , title =

    work page

  14. [14]

    I , year =

    Machine Learning: An Artificial Intelligence Approach, Vol. I , year =

    work page

  15. [15]

    T. M. Mitchell , institution =. The Need for Biases in Learning Generalizations , year =

    work page

  16. [16]

    Newell and P

    A. Newell and P. S. Rosenbloom , booktitle =. Mechanisms of Skill Acquisition and the Law of Practice , year =

    work page

  17. [17]

    Xiaoyu Li and Yitian Zhang and Guodong Long and Yupeng Hu and Wenpeng Lu and Meng Chen and Chengqi Zhang and Yongshun Gong , title =

    work page

  18. [18]

    NeurIPS , year =

    Lei Bai and Lina Yao and Can Li and Xianzhi Wang and Can Wang , title =. NeurIPS , year =

    work page

  19. [19]

    Yudong Zhang and Pengkun Wang and Binwu Wang and Xu Wang and Zhe Zhao and Zhengyang Zhou and Lei Bai and Yang Wang , title =

    work page

  20. [20]

    Shengnan Guo and Youfang Lin and Ning Feng and Chao Song and Huaiyu Wan , title =

    work page

  21. [21]

    KDD , pages =

    Ting Li and Junbo Zhang and Kainan Bao and Yuxuan Liang and Yexin Li and Yu Zheng , title =. KDD , pages =

    work page

  22. [22]

    KDD , pages =

    Tengfei Lyu and Weijia Zhang and Jinliang Deng and Hao Liu , title =. KDD , pages =

    work page

  23. [23]

    Jensen and Xueqi Cheng , title =

    Zezhi Shao and Fei Wang and Yongjun Xu and Wei Wei and Chengqing Yu and Zhao Zhang and Di Yao and Tao Sun and Guangyin Jin and Xin Cao and Gao Cong and Christian S. Jensen and Xueqi Cheng , title =. IEEE Trans. Knowl. Data Eng. , volume =

    work page

  24. [24]

    Jindong Han and Weijia Zhang and Hao Liu and Tao Tao and Naiqiang Tan and Hui Xiong , title =. Proc. VLDB Endow. , volume =

    work page

  25. [25]

    Proceedings of the AAAI Conference on Artificial Intelligence , pages=

    Causal conditional hidden Markov model for multimodal traffic prediction , author=. Proceedings of the AAAI Conference on Artificial Intelligence , pages=

    work page

  26. [26]

    Fleet and Mohammad Norouzi and Tim Salimans , title =

    Jonathan Ho and Chitwan Saharia and William Chan and David J. Fleet and Mohammad Norouzi and Tim Salimans , title =. J. Mach. Learn. Res. , volume =

    work page

  27. [27]

    CIKM , pages =

    Xiangyu Zhao and Jiliang Tang , title =. CIKM , pages =

    work page

  28. [28]

    NeurIPS , year =

    Xu Liu and Yutong Xia and Yuxuan Liang and Junfeng Hu and Yiwei Wang and Lei Bai and Chao Huang and Zhenguang Liu and Bryan Hooi and Roger Zimmermann , title =. NeurIPS , year =

    work page

  29. [29]

    ICLR , year =

    Yaguang Li and Rose Yu and Cyrus Shahabi and Yan Liu , title =. ICLR , year =

    work page

  30. [30]

    IEEE Trans

    Zulong Diao and Xin Wang and Dafang Zhang and Gaogang Xie and Jianguo Chen and Changhua Pei and Xuying Meng and Kun Xie and Guangxing Zhang , title =. IEEE Trans. Mob. Comput. , volume =

    work page

  31. [31]

    ICML , pages =

    Shiyong Lan and Yitong Ma and Weikang Huang and Wenwu Wang and Hongyu Yang and Pyang Li , title =. ICML , pages =

    work page

  32. [32]

    AAAI , pages =

    Sicheng He and Junzhong Ji and Minglong Lei , title =. AAAI , pages =

    work page

  33. [33]

    AAAI , pages =

    Chuanpan Zheng and Xiaoliang Fan and Cheng Wang and Jianzhong Qi , title =. AAAI , pages =

    work page

  34. [34]

    IJCAI , pages =

    Zonghan Wu and Shirui Pan and Guodong Long and Jing Jiang and Chengqi Zhang , title =. IJCAI , pages =

    work page

  35. [35]

    AAAI , pages =

    Kan Guo and Yongli Hu and Yanfeng Sun and Sean Qian and Junbin Gao and Baocai Yin , title =. AAAI , pages =

    work page

  36. [36]

    CIKM , pages =

    Qian Ma and Zijian Zhang and Xiangyu Zhao and Haoliang Li and Hongwei Zhao and Yiqi Wang and Zitao Liu and Wanyu Wang , title =. CIKM , pages =

    work page

  37. [37]

    KDD , pages =

    Zheng Dong and Renhe Jiang and Haotian Gao and Hangchen Liu and Jinliang Deng and Qingsong Wen and Xuan Song , title =. KDD , pages =

    work page

  38. [38]

    Hierarchical Spatio-Temporal Graph Neural Networks for Pandemic Forecasting , booktitle =

    Yihong Ma and Patrick G. Hierarchical Spatio-Temporal Graph Neural Networks for Pandemic Forecasting , booktitle =

    work page

  39. [39]

    KDD , pages =

    Ning Wu and Wayne Xin Zhao and Jingyuan Wang and Dayan Pan , title =. KDD , pages =

    work page

  40. [40]

    Denton and Seyed Kamyar Seyed Ghasemipour and Raphael Gontijo Lopes and Burcu Karagol Ayan and Tim Salimans and Jonathan Ho and David J

    Chitwan Saharia and William Chan and Saurabh Saxena and Lala Li and Jay Whang and Emily L. Denton and Seyed Kamyar Seyed Ghasemipour and Raphael Gontijo Lopes and Burcu Karagol Ayan and Tim Salimans and Jonathan Ho and David J. Fleet and Mohammad Norouzi , title =. NeurIPS , year =

    work page

  41. [41]

    Lu and Pedram Hassanzadeh and Michael Maire and Rebecca Willett , title =

    Ruoxi Jiang and Xiao Zhang and Karan Jakhar and Peter Y. Lu and Pedram Hassanzadeh and Michael Maire and Rebecca Willett , title =. NeurIPS , year =

    work page

  42. [42]

    ICDE , pages =

    Nan Jiang and Haitao Yuan and Jianing Si and Minxiao Chen and Shangguang Wang , title =. ICDE , pages =

    work page

  43. [43]

    IJCAI , pages =

    Zhongjian Lv and Jiajie Xu and Kai Zheng and Hongzhi Yin and Pengpeng Zhao and Xiaofang Zhou , title =. IJCAI , pages =

    work page

  44. [44]

    ICDE , pages =

    Shengnan Guo and Youfang Lin and Letian Gong and Chenyu Wang and Zeyu Zhou and Zekai Shen and Yiheng Huang and Huaiyu Wan , title =. ICDE , pages =

    work page

  45. [45]

    Zheng Fang and Qingqing Long and Guojie Song and Kunqing Xie , title =

    work page

  46. [46]

    Transportation research part C: emerging technologies , volume=

    Percolation-based dynamic perimeter control for mitigating congestion propagation in urban road networks , author=. Transportation research part C: emerging technologies , volume=. 2022 , publisher=

    work page 2022

  47. [47]

    RPMixer: Shaking Up Time Series Forecasting with Random Projections for Large Spatial-Temporal Data , booktitle =

    Chin. RPMixer: Shaking Up Time Series Forecasting with Random Projections for Large Spatial-Temporal Data , booktitle =

    work page

  48. [48]

    Jensen , title =

    Zezhi Shao and Zhao Zhang and Wei Wei and Fei Wang and Yongjun Xu and Xin Cao and Christian S. Jensen , title =. Proc

    work page

  49. [49]

    Fuxian Li and Jie Feng and Huan Yan and Guangyin Jin and Fan Yang and Funing Sun and Depeng Jin and Yong Li , title =

    work page

  50. [50]

    IEEE Trans

    Hao Shu and Hailin Wang and Jiangjun Peng and Deyu Meng , title =. IEEE Trans. Intell. Transp. Syst. , volume =

    work page

  51. [51]

    KDD , pages =

    Tong Nie and Guoyang Qin and Wei Ma and Yuewen Mei and Jian Sun , title =. KDD , pages =

    work page

  52. [52]

    ISDA , pages =

    Pasquale De Meo and Emilio Ferrara and Giacomo Fiumara and Alessandro Provetti , title =. ISDA , pages =

    work page

  53. [53]

    IEEE Trans

    Ling Chen and Donghui Chen and Zongjiang Shang and Binqing Wu and Cen Zheng and Bo Wen and Wei Zhang , title =. IEEE Trans. Knowl. Data Eng. , volume =

    work page

  54. [54]

    Yu , title =

    Senzhang Wang and Meiyue Zhang and Hao Miao and Zhaohui Peng and Philip S. Yu , title =. ACM Trans. Intell. Syst. Technol. , volume =

    work page

  55. [55]

    IEEE Trans

    Junbo Zhang and Yu Zheng and Junkai Sun and Dekang Qi , title =. IEEE Trans. Knowl. Data Eng. , volume =

    work page

  56. [56]

    AAAI , pages =

    Renhe Jiang and Zhaonan Wang and Jiawei Yong and Puneet Jeph and Quanjun Chen and Yasumasa Kobayashi and Xuan Song and Shintaro Fukushima and Toyotaro Suzumura , title =. AAAI , pages =

    work page

  57. [57]

    KDD , pages =

    Zonghan Wu and Shirui Pan and Guodong Long and Jing Jiang and Xiaojun Chang and Chengqi Zhang , title =. KDD , pages =

    work page

  58. [58]

    Salim , title =

    Wei Shao and Zhiling Jin and Shuo Wang and Yufan Kang and Xiao Xiao and Hamid Menouar and Zhaofeng Zhang and Junshan Zhang and Flora D. Salim , title =. IJCAI , pages =

    work page

  59. [59]

    KDD , pages =

    Yuchen Fang and Yuxuan Liang and Bo Hui and Zezhi Shao and Liwei Deng and Xu Liu and Xinke Jiang and Kai Zheng , title =. KDD , pages =

    work page

  60. [60]

    AAAI , pages =

    Jiawei Jiang and Chengkai Han and Wayne Xin Zhao and Jingyuan Wang , title =. AAAI , pages =

    work page

  61. [61]

    AAAI , pages =

    Weiyang Kong and Ziyu Guo and Yubao Liu , title =. AAAI , pages =

    work page

  62. [62]

    Ng and Michael I

    Andrew Y. Ng and Michael I. Jordan and Yair Weiss , title =. NIPS , pages =

    work page

  63. [63]

    NeurIPS , year =

    Minhao Liu and Ailing Zeng and Muxi Chen and Zhijian Xu and Qiuxia Lai and Lingna Ma and Qiang Xu , title =. NeurIPS , year =

    work page

  64. [64]

    IEEE Trans

    Shengnan Guo and Youfang Lin and Shijie Li and Zhaoming Chen and Huaiyu Wan , title =. IEEE Trans. Intell. Transp. Syst. , volume =

    work page

  65. [65]

    CIKM , pages =

    Hangchen Liu and Zheng Dong and Renhe Jiang and Jiewen Deng and Jinliang Deng and Quanjun Chen and Xuan Song , title =. CIKM , pages =

    work page

  66. [66]

    IEEE Intell

    Ge Guo and Wei Yuan and Jinyuan Liu and Yisheng Lv and Wei Liu , title =. IEEE Intell. Transp. Syst. Mag. , volume =

    work page

  67. [67]

    Spatio-Temporal Predictive Modeling Techniques for Different Domains: a Survey , journal =

    Rahul Kumar and Manish Bhanu and Jo. Spatio-Temporal Predictive Modeling Techniques for Different Domains: a Survey , journal =

    work page

  68. [68]

    CIKM , pages =

    Cheonbok Park and Chunggi Lee and Hyojin Bahng and Yunwon Tae and Seungmin Jin and Kihwan Kim and Sungahn Ko and Jaegul Choo , title =. CIKM , pages =

    work page

  69. [69]

    IJCAI , pages =

    Bing Yu and Haoteng Yin and Zhanxing Zhu , title =. IJCAI , pages =

    work page

  70. [70]

    ICDE , pages =

    Zhonghang Li and Chao Huang and Lianghao Xia and Yong Xu and Jian Pei , title =. ICDE , pages =

    work page

  71. [71]

    CIKM , pages =

    Zezhi Shao and Zhao Zhang and Fei Wang and Wei Wei and Yongjun Xu , title =. CIKM , pages =

    work page

  72. [72]

    Yu , title =

    Chuanpan Zheng and Xiaoliang Fan and Shirui Pan and Haibing Jin and Zhaopeng Peng and Zonghan Wu and Cheng Wang and Philip S. Yu , title =. IEEE Trans. Knowl. Data Eng. , volume =

    work page

  73. [73]

    AAAI , pages =

    Zetao Li and Zheng Hu and Peng Han and Yu Gu and Shimin Cai , title =. AAAI , pages =

    work page

  74. [74]

    AAAI , pages =

    Junbo Zhang and Yu Zheng and Dekang Qi , title =. AAAI , pages =

    work page

  75. [75]

    AAAI , pages =

    Jiahao Ji and Jingyuan Wang and Chao Huang and Junjie Wu and Boren Xu and Zhenhe Wu and Junbo Zhang and Yu Zheng , title =. AAAI , pages =

    work page

  76. [76]

    ICDE , pages =

    Yuchen Fang and Yanjun Qin and Haiyong Luo and Fang Zhao and Bingbing Xu and Liang Zeng and Chenxing Wang , title =. ICDE , pages =

    work page

  77. [77]

    NeurIPS , year=

    Less but more: Linear adaptive graph learning empowering spatiotemporal forecasting , author=. NeurIPS , year=

    work page

  78. [78]

    NeurIPS , year=

    How Different from the Past? Spatio-Temporal Time Series Forecasting with Self-Supervised Deviation Learning , author=. NeurIPS , year=

    work page

  79. [79]

    Ziye Wang and Yiran Qin and Lin Zeng and Ruimao Zhang , title =

    work page

  80. [80]

    Liangzhe Han and Bowen Du and Leilei Sun and Yanjie Fu and Yisheng Lv and Hui Xiong , title =

    work page

Showing first 80 references.