Damba-ST: Domain-Adaptive Mamba for Efficient Urban Spatio-Temporal Prediction
Rui An, Yifeng Zhang, Ziran Liang, Wenqi Fan, Yuxuan Liang, Xuequn Shang, Qing Li
TL;DR
This work tackles cross-city urban spatio-temporal forecasting under domain heterogeneity and Transformer-era computational constraints. It introduces Damba-ST, a Domain-Adaptive Mamba backbone built around a Domain-Adaptive State Space Model (DASSM) with Domain Adapters, and three complementary views (Spatial, Temporal, ST-delay) processed via Intra-Domain Scanning and Cross-Domain Adaptation. The model learns domain-specific discriminative patterns while aligning cross-domain commonalities, achieving linear complexity in sequence length and strong zero-shot generalization across regions, cities, and tasks. Empirical results show state-of-the-art performance and practical deployment benefits, including fast inference on real-world data, enabling robust, plug-and-play urban forecasting without extensive retraining.
Abstract
Training urban spatio-temporal foundation models that generalize well across diverse regions and cities is critical for deploying urban services in unseen or data-scarce regions. Recent studies have typically focused on fusing cross-domain spatio-temporal data to train unified Transformer-based models. However, these models suffer from quadratic computational complexity and high memory overhead, limiting their scalability and practical deployment. Inspired by the efficiency of Mamba, a state space model with linear time complexity, we explore its potential for efficient urban spatio-temporal prediction. However, directly applying Mamba as a spatio-temporal backbone leads to negative transfer and severe performance degradation. This is primarily due to spatio-temporal heterogeneity and the recursive mechanism of Mamba's hidden state updates, which limit cross-domain generalization. To overcome these challenges, we propose Damba-ST, a novel domain-adaptive Mamba-based model for efficient urban spatio-temporal prediction. Damba-ST retains Mamba's linear complexity advantage while significantly enhancing its adaptability to heterogeneous domains. Specifically, we introduce two core innovations: (1) a domain-adaptive state space model that partitions the latent representation space into a shared subspace for learning cross-domain commonalities and independent, domain-specific subspaces for capturing intra-domain discriminative features; (2) three distinct Domain Adapters, which serve as domain-aware proxies to bridge disparate domain distributions and facilitate the alignment of cross-domain commonalities. Extensive experiments demonstrate the generalization and efficiency of Damba-ST. It achieves state-of-the-art performance on prediction tasks and demonstrates strong zero-shot generalization, enabling seamless deployment in new urban environments without extensive retraining or fine-tuning.