---
title: "Addressing Downward Memory Loss in Hierarchical GNN Forecasters Through Memory-Buffered Decoding"
authors:
  - "Thomas Bailie"
  - "S. Karthik Mukkavilli"
  - "Varvara Vetrova"
  - "Yun Sing Koh"
doi: "10.24963/ijcai.2026/462"
url: "https://ijcai.org/proceedings/2026/462"
pdf: "https://ijcai.org/proceedings/2026/0462.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "4152-4160"
track: "Main Track"
keywords:
  - "Machine Learning: Time series and data streams"
  - "Machine Learning: Sequence and graph learning"
  - "Machine Learning: Multi-view learning"
  - "Machine Learning: Learning graphical models"
---

# Addressing Downward Memory Loss in Hierarchical GNN Forecasters Through Memory-Buffered Decoding

Thomas Bailie, S. Karthik Mukkavilli, Varvara Vetrova, Yun Sing Koh

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 4152-4160. DOI: [10.24963/ijcai.2026/462](https://doi.org/10.24963/ijcai.2026/462)

## Abstract

Accurate spatio-temporal forecasting requires modeling interactions across multiple spatial and temporal scales. Existing Graph Neural Network (GNN) forecasters primarily operate at a single local scale, limiting their ability to capture global processes that govern system dynamics. Hierarchical GNNs (HGNNs) aim to address this limitation by learning multiscale representations, but in practice, they often fail to preserve global information during coarse-to-fine propagation. We identify this limitation as downward memory loss, where global trends learned at coarse resolutions diminish before influencing fine-grained predictions, leading to misaligned local dynamics. We propose HiGFlow, an HGNN forecaster that explicitly preserves multiscale trends through a self-updating memory buffer that integrates into the coarse-to-fine information flow. This design maintains global contextual signals as an inductive bias throughout hierarchical decoding. We provide a theoretical analysis indicating that when decoding, HiGFlow preserves multiscale trends where residual-based architectures may fail to do so. Empirically, HiGFlow achieves substantially lower MAE and RMSE than state-of-the-art forecasting models across multiple benchmark datasets, demonstrating the importance of explicit global memory in multiscale spatio-temporal forecasting. Our implementation is available at https://github.com/TB862/HiGFlow.