Reducing Smoothness with Expressive Memory Enhanced Hierarchical Graph Neural Networks
Not on the map. Why?
Abstract (arXiv)
Graphical forecasting models learn the structure of time series data via projecting onto a graph, with recent techniques capturing spatial-temporal associations between variables via edge weights. Hierarchical variants offer a distinct advantage by analysing the time series across multiple resolutions, making them particularly effective in tasks like global weather forecasting, where low-resolution variable interactions are significant. A critical challenge in hierarchical models is information loss during forward or backward passes through the hierarchy. We propose the Hierarchical Graph Flow (HiGFlow) network, which introduces a memory buffer variable of dynamic size to store previously seen information across variable resolutions. We theoretically show two key results: HiGFlow reduces smoothness when mapping onto new feature spaces in the hierarchy and non-strictly enhances the utility of message-passing by improving Weisfeiler-Lehman (WL) expressivity. Empirical results demonstrate that HiGFlow outperforms state-of-the-art baselines, including transformer models, by at least an average of 6.1% in MAE and 6.2% in RMSE. Code is available at https://github.com/TB862/ HiGFlow.git.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 28 of its 41 references that are in the map. First posted on arXiv 1 Apr 2025.
Builds on
28 of 41 references.
References · 28
- 2017Attention Is All You Needcited by 60,361
- 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
- 2017Graph Attention Networkscited by 8,923
- 2017Inductive Representation Learning on Large Graphscited by 6,320
- 2018How Powerful are Graph Neural Networks?cited by 4,393
- 2020Fourier Neural Operator for Parametric Partial Differential Equationscited by 2,908
- 2014Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modelingcited by 2,367
- 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 2,039
- 2021Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecastingcited by 1,676
- 2022FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecastingcited by 1,210
- 2023iTransformer: Inverted Transformers Are Effective for Time Series Forecastingcited by 931
- 2018Hierarchical Graph Representation Learning with Differentiable Poolingcited by 787
16 more
- 2020On the Bottleneck of Graph Neural Networks and its Practical Implicationscited by 631
- 2022Scientific Machine Learning through Physics-Informed Neural Networks: Where we are and What's nextcited by 608
- 2019Self-Attention Graph Poolingcited by 421
- 2021Spectral Temporal Graph Neural Network for Multivariate Time-series Forecastingcited by 299
- 2023A Survey on Oversmoothing in Graph Neural Networkscited by 247
- 2023PDE-Refiner: Achieving Accurate Long Rollouts with Neural PDE Solverscited by 216
- 2022Transformer Quality in Linear Timecited by 191
- 2023Frequency-domain MLPs are More Effective Learners in Time Series Forecastingcited by 172
- 2023On Over-Squashing in Message Passing Neural Networks: The Impact of Width, Depth, and Topologycited by 166
- 2023FourierGNN: Rethinking Multivariate Time Series Forecasting from a Pure Graph Perspectivecited by 127
- 2019Hierarchical Graph Pooling with Structure Learningcited by 81
- 2018Multilevel Wavelet Decomposition Network for Interpretable Time Series Analysiscited by 61
- 2023The expressive power of pooling in Graph Neural Networkscited by 35
- 2024Probabilistic Weather Forecasting with Hierarchical Graph Neural Networkscited by 30
- 2023Graph-based Time Series Clustering for End-to-End Hierarchical Forecastingcited by 17
- 2020Hierarchical Graph Capsule Networkcited by 11
Built on by
Not cited yet.