Pith. sign in

REVIEW 22 cited by

Spatial-Temporal Transformer Networks for Traffic Flow Forecasting

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2001.02908 v2 pith:AX27FERP submitted 2020-01-09 eess.SP cs.LG

Spatial-Temporal Transformer Networks for Traffic Flow Forecasting

Mingxing Xu , Wenrui Dai , Chunmiao Liu , Xing Gao , Weiyao Lin , Guo-Jun Qi , Hongkai Xiong This is my paper
classification eess.SP cs.LG
keywords trafficdependenciesforecastingspatial-temporalmodelspatialtransformerflows
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Traffic forecasting has emerged as a core component of intelligent transportation systems. However, timely accurate traffic forecasting, especially long-term forecasting, still remains an open challenge due to the highly nonlinear and dynamic spatial-temporal dependencies of traffic flows. In this paper, we propose a novel paradigm of Spatial-Temporal Transformer Networks (STTNs) that leverages dynamical directed spatial dependencies and long-range temporal dependencies to improve the accuracy of long-term traffic forecasting. Specifically, we present a new variant of graph neural networks, named spatial transformer, by dynamically modeling directed spatial dependencies with self-attention mechanism to capture realtime traffic conditions as well as the directionality of traffic flows. Furthermore, different spatial dependency patterns can be jointly modeled with multi-heads attention mechanism to consider diverse relationships related to different factors (e.g. similarity, connectivity and covariance). On the other hand, the temporal transformer is utilized to model long-range bidirectional temporal dependencies across multiple time steps. Finally, they are composed as a block to jointly model the spatial-temporal dependencies for accurate traffic prediction. Compared to existing works, the proposed model enables fast and scalable training over a long range spatial-temporal dependencies. Experiment results demonstrate that the proposed model achieves competitive results compared with the state-of-the-arts, especially forecasting long-term traffic flows on real-world PeMS-Bay and PeMSD7(M) datasets.

discussion (0)

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

Forward citations

Cited by 22 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. RotVLA: Rotational Latent Action for Vision-Language-Action Model

    cs.RO 2026-05 unverdicted novelty 7.0

    RotVLA models latent actions as continuous SO(n) rotations with triplet-frame supervision and flow-matching to reach 98.2% success on LIBERO and 89.6%/88.5% on RoboTwin2.0 using a 1.7B-parameter model.

  2. AirQualityBench: A Realistic Evaluation Benchmark for Global Air Quality Forecasting

    cs.AI 2026-05 accept novelty 7.0

    AirQualityBench is a realistic global benchmark using hourly data from 3720 stations across 2021-2025 for six pollutants, preserving native missingness masks and evaluating on inverse-transformed physical scales.

  3. Incident-Guided Spatiotemporal Traffic Forecasting

    cs.LG 2026-01 unverdicted novelty 7.0

    IGSTGNN adds incident-context spatial fusion and temporal impact decay modules to model how events alter traffic patterns, achieving state-of-the-art results on a new time-aligned incident-traffic dataset.

  4. Revisiting the Adversarial Robustness of Graph-Based Traffic Forecasting

    cs.CR 2026-07 conditional novelty 6.0

    A physics-aware, localized attack defeats norm-bounded adversarial training in traffic forecasting, and a physics-informed detection feature (VetTraffic) outperforms attack-specific hardening on 13/15 settings.

  5. A2TTA: Anchored-and-Agile Test-Time Adaptation for Evolving Traffic Sensor Networks

    cs.LG 2026-07 conditional novelty 6.0

    A2TTA keeps a frozen traffic-forecasting backbone and adapts an expandable FiLM calibrator with a persistent global state and a single-use local clone, cutting MAE on ten evolving sensor networks.

  6. Do We Really Need Adaptive Global Spatial Attention for Traffic Forecasting?

    cs.AI 2026-07 conditional novelty 6.0

    In a controlled comparison across six traffic benchmarks, standard spatial attention did not consistently beat a simple node-mean mixer, suggesting adaptive global attention is not essential.

  7. DREAMSTEER: Latent World Models Can Steer VLA Policies During Deployment Without Any Finetuning

    cs.RO 2026-07 conditional novelty 6.0

    A frozen VLA plus latent world-model rollouts and a value model can raise real-robot OOD manipulation success from 23.75% to 66.25% without any target-environment finetuning.

  8. Flowing With Purpose: Latent Action Guided Flow Matching Policies For Robotic Manipulation

    cs.RO 2026-06 unverdicted novelty 6.0

    LAFM adapts the source distribution in flow matching policies via a latent action model to better match fragmented robotic action spaces, claiming 23.4% higher real-world success and 10.4% on LIBERO-90 while beating l...

  9. Hierarchical Flow Decomposition for Turning Movement Prediction at Signalized Intersections

    cs.LG 2026-04 unverdicted novelty 6.0

    HFD-TM predicts turning movements with 2.49 MAE by hierarchically decomposing corridor flows and enforcing conservation, outperforming Transformer and GRU baselines on six months of Nashville LiDAR data.

  10. UNICA: A Unified Neural Framework for Controllable 3D Avatars

    cs.CV 2026-04 unverdicted novelty 6.0

    UNICA unifies motion planning, rigging, physical simulation, and rendering into a single skeleton-free neural framework that produces next-frame 3D avatar geometry from action inputs and renders it with Gaussian splatting.

  11. Co-Evolving Latent Action World Models

    cs.LG 2025-10 unverdicted novelty 6.0

    CoLA-World jointly trains latent action models and world models with a warm-up phase to achieve co-evolution, matching or exceeding prior two-stage methods in video simulation quality and visual planning performance.

  12. villa-X: Enhancing Latent Action Modeling in Vision-Language-Action Models

    cs.RO 2025-07 unverdicted novelty 6.0

    villa-X enhances latent action modeling in VLA models to support zero-shot action planning for unseen robot embodiments and open-vocabulary instructions, yielding better manipulation results in simulation and real-wor...

  13. UniVLA: Learning to Act Anywhere with Task-centric Latent Actions

    cs.RO 2025-05 unverdicted novelty 6.0

    UniVLA trains cross-embodiment vision-language-action policies from unlabeled videos via a latent action model in DINO space, beating OpenVLA on benchmarks with 1/20th pretraining compute and 1/10th downstream data.

  14. AgiBot World Colosseo: A Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systems

    cs.RO 2025-03 unverdicted novelty 6.0

    AgiBot World supplies over 1 million trajectories enabling GO-1 to deliver 30% average gains over Open X-Embodiment and over 60% success on complex dexterous tasks while open-sourcing everything.

  15. Do We Really Need Adaptive Global Spatial Attention for Traffic Forecasting?

    cs.AI 2026-07 accept novelty 5.5

    Uniform full-range mean broadcasting matches standard spatial attention on six traffic benchmarks (0.14% mean MAE gap) while cutting node mixing cost from O(N²) to O(N).

  16. MambaLSTM: A Spatio-Temporal Framework for Enhanced Traffic Accident Risk Prediction

    cs.LG 2026-07 conditional novelty 5.0

    MambaLSTM, a Mamba-plus-LSTM spatiotemporal network, reports F1-score gains of roughly 2–5 points over baselines for next-step traffic accident risk prediction on NYC, Chicago, and US-Accidents data.

  17. Efficient Prompt Learning for Traffic Forecasting

    cs.LG 2026-05 unverdicted novelty 5.0

    SimpleST is a model-agnostic prompt tuning framework that lets pre-trained spatio-temporal GNNs adapt to distribution shifts in traffic data while keeping all original model weights fixed.

  18. Eliminating Propagation Delay: Attention-Based Spatial-Temporal Fusion Graph Convolution Network for Traffic Flow Prediction

    cs.LG 2026-07 conditional novelty 4.0

    A graph model that injects 5–10 minute DTW delay edges, spectral macro clusters, and masked temporal attention beats common traffic baselines on five PeMS/LargeST datasets.

  19. STAGformer: A Spatio-temporal Agent Graph Transformer for Micro Mobility Demand Forecasting

    cs.LG 2026-07 conditional novelty 4.0

    STAGformer forecasts bike-station demand with linear-complexity spatial-temporal agent attention and reports lower RMSE/MAE than listed baselines on NYC and Chicago data.

  20. Dimensional Balance Improves Large Scale Spatiotemporal Prediction Performance

    cs.LG 2026-05 unverdicted novelty 4.0

    A scalable framework harmonizes spatial and temporal representations via low-rank spatial compression and extended temporal horizons to reduce prediction uncertainty in large-scale spatiotemporal tasks.

  21. Deep Spatio-Temporal Neural Network for Air Quality Reanalysis

    cs.LG 2025-02 unverdicted novelty 3.0

    AQ-Net combines LSTM-attention for time and neural kNN for space to reanalyze PM2.5 at monitored and unmonitored stations using 2013-2017 northern China data.

  22. EMAGN: Efficient Multi-Attention Graph Network via Learned Clustering for Scalable Traffic Forecasting

    cs.LG 2026-07 reject novelty 2.0

    EMAGN compresses spatial attention keys/values into M learned summaries, matching GMAN within ~3% MAE at 60 min while using 58% less memory and enabling 16-head attention on an 11 GB GPU.