LLMODE: Aligning ODEs with LLMs via Gated Token Injection for Irregular Spatio-Temporal Forecasting
Machine LearningArtificial Intelligence
Summary
The authors developed LLMODE, a method that helps large language models make predictions from data collected at uneven times or places, like irregular traffic or weather readings. They use special techniques to turn irregular data into smooth, continuous information and then compress it so the model can understand it efficiently. Their approach combines different types of memory to give the model a clearer picture, even when data is sparse or complex. Tests showed that their method works well on various urban and physical datasets, and it can even predict in new locations it hasn't seen before without extra training.
Authors
Di Zhang, Jingyang Zhang, Ziqian Wang, Chi Zhang, Yikun Ban, Ziwei Zhang, Ruijie Wang
Abstract
Large language models (LLMs) have shown promise for spatio-temporal forecasting, but existing approaches often rely on regularly sampled token sequences and struggle with irregular observations because of temporal asynchrony, representation-space misalignment, and limited context windows. We propose LLMODE, a token-efficient framework for irregular spatio-temporal forecasting with a frozen LLM backbone. LLMODE first uses a graph-aware ODE encoder to reconstruct irregular graph observations as a continuous-time latent trajectory. A Fixed-Budget Perceiver Resampler then compresses this variable-length trajectory into a fixed number of dynamic memory tokens. In parallel, compact statistical descriptors are encoded and resampled into context memory tokens. A dual-source gated cross-attention module injects both memories into the frozen LLM, enabling controlled utilization of external spatio-temporal evidence. Experiments on three real-world urban datasets and two physical-dynamics benchmarks show competitive overall performance, with clearer advantages under sparse or dynamically complex irregular sampling. Additional evaluations on unseen urban regions further demonstrate strong zero-shot generalization without adaptation.