Pith. sign in

REVIEW 5 cited by

STGformer: Efficient Spatiotemporal Graph Transformer for Traffic Forecasting

Not yet reviewed by Pith; the record is open.

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 2410.00385 v2 pith:4ZEHL66E submitted 2024-10-01 cs.LG cs.AIcs.DB

STGformer: Efficient Spatiotemporal Graph Transformer for Traffic Forecasting

classification cs.LG cs.AIcs.DB
keywords spatiotemporalstgformertrafficcomputationalforecastinggraphefficientapproaches
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Traffic forecasting is a cornerstone of smart city management, enabling efficient resource allocation and transportation planning. Deep learning, with its ability to capture complex nonlinear patterns in spatiotemporal (ST) data, has emerged as a powerful tool for traffic forecasting. While graph neural networks (GCNs) and transformer-based models have shown promise, their computational demands often hinder their application to real-world road networks, particularly those with large-scale spatiotemporal interactions. To address these challenges, we propose a novel spatiotemporal graph transformer (STGformer) architecture. STGformer effectively balances the strengths of GCNs and Transformers, enabling efficient modeling of both global and local traffic patterns while maintaining a manageable computational footprint. Unlike traditional approaches that require multiple attention layers, STG attention block captures high-order spatiotemporal interactions in a single layer, significantly reducing computational cost. In particular, STGformer achieves a 100x speedup and a 99.8\% reduction in GPU memory usage compared to STAEformer during batch inference on a California road graph with 8,600 sensors. We evaluate STGformer on the LargeST benchmark and demonstrate its superiority over state-of-the-art Transformer-based methods such as PDFormer and STAEformer, which underline STGformer's potential to revolutionize traffic forecasting by overcoming the computational and memory limitations of existing approaches, making it a promising foundation for future spatiotemporal modeling tasks.

discussion (0)

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

Forward citations

Cited by 5 Pith papers

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

  1. K-STEMIT: Knowledge-Informed Spatio-Temporal Efficient Multi-Branch Graph Neural Network for Subsurface Stratigraphy Thickness Estimation from Radar Data

    cs.LG 2026-04 unverdicted novelty 6.0

    K-STEMIT reduces RMSE by 21% for subsurface stratigraphy thickness estimation from radar data via a knowledge-informed spatio-temporal GNN with adaptive feature fusion and physical priors from the MAR weather model.

  2. Efficient Traffic Prediction at Scale: A Systematic Study of STGCN Architectural Depth

    cs.LG 2026-06 unverdicted novelty 5.0

    Systematic experiments on four traffic datasets find that a 1-block STGCN achieves optimal short-term (10 min) prediction on three datasets with only marginal longer-horizon degradation and 61% lower CPU latency than ...

  3. K-STEMIT: Knowledge-Informed Spatio-Temporal Efficient Multi-Branch Graph Neural Network for Subsurface Stratigraphy Thickness Estimation from Radar Data

    cs.LG 2026-04 conditional novelty 5.0

    Adding MAR climate-model features plus a multi-branch GraphSAGE/temporal-convolution design with adaptive fusion reduces deep-ice-layer thickness RMSE by 21.01% over a no-knowledge multi-branch baseline on the SRED Gr...

  4. 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.

  5. 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.