Just for FUNS: LLM-Guided Spatio-Temporal Graph Node Generation for Forecasting Unobserved Node States

2026-09-24

AI Summary

This study addresses the challenge of inferring unobserved node states in spatiotemporal prediction caused by insufficient sensor coverage. To this end, we propose GenST, a novel framework that pioneers the use of large language models (LLMs) as semantic bridges to compensate for missing signals. Specifically, GenST reformulates prediction as a conditional generation task: it extracts semantic features via LLMs and integrates a spatiotemporal variational autoencoder with a generative Transformer to reconstruct future states of unobserved nodes under multimodal conditioning. Extensive experiments demonstrate that GenST significantly outperforms existing baselines across multiple datasets, effectively mitigating data sparsity while achieving high-accuracy zero-shot prediction.

Abstract

Spatio-temporal forecasting is a cornerstone of logistics, urban planning, and intelligent transportation systems. However, constrained by deployment costs and maintenance resources, sensor networks often lack comprehensive spatial coverage, rendering Forecast Unobserved Node States (FUNS) a critical yet formidable challenge. Conventional models rely on historical observations and typically falter when encountering nodes without prior records. To address this, we redefine the problem as a conditional generation task on spatio-temporal graphs and propose GenST, a framework that introduces Large Language Models (LLMs) as a semantic bridge, leveraging a pre-trained LLM fine-tuned to extract rich semantic features from node descriptions, such as functional zones and road network structures, to compensate for missing spatio-temporal signals. Specifically, we design a two-stage generative architecture: a Spatio-Temporal VAE first compresses spatio-temporal dynamics into a latent space, followed by a Generative Transformer (GenT) that reconstructs the future states of unobserved nodes from noise, guided by multi-modal conditions including semantics, geographic coordinates, and neighborhood contexts. Experiments on six traffic and two non-traffic datasets show GenST significantly outperforms existing baselines in zero-shot prediction tasks, demonstrating the practical potential of semantic-guided generation for mitigating spatio-temporal data sparsity.

Problem

• spatio-temporal forecasting
• unobserved node states
• zero-shot prediction
• data sparsity
• sensor networks

Innovation

• Spatio-temporal forecasting
• Large Language Models
• Conditional generation
• Variational Autoencoder
• Zero-shot prediction