Search

Spatiotemporal time series forecasting plays a key role in a wide range of real-world applications. While significant progress has been made in this area, fully capturing and leveraging spatiotemporal heterogeneity remains a fundamental challenge. Therefore, we propose a novel Heterogeneity-Informed Meta-Parameter Learning scheme. Specifically, our approach implicitly captures spatiotemporal heterogeneity through learning spatial and temporal embeddings, which can be viewed as a clustering process. Then, a novel spatiotemporal meta-parameter learning paradigm is proposed to learn spatiotemporal-specific parameters from meta-parameter pools, which is informed by the captured heterogeneity. Based on these ideas, we develop a Heterogeneity-Informed Spatiotemporal Meta-Network (HimNet) for spatiotemporal time series forecasting. Extensive experiments on five widely-used benchmarks demonstrate our method achieves state-of-the-art performance while exhibiting superior interpretability. Our code is available at https://github.com/XDZhelheim/HimNet., Comment: Accepted by KDD'24 Research Track

Spatiotemporal forecasting techniques are significant for various domains such as transportation, energy, and weather. Accurate prediction of spatiotemporal series remains challenging due to the complex spatiotemporal heterogeneity. In particular, current end-to-end models are limited by input length and thus often fall into spatiotemporal mirage, i.e., similar input time series followed by dissimilar future values and vice versa. To address these problems, we propose a novel self-supervised pre-training framework Spatial-Temporal-Decoupled Masked Pre-training (STD-MAE) that employs two decoupled masked autoencoders to reconstruct spatiotemporal series along the spatial and temporal dimensions. Rich-context representations learned through such reconstruction could be seamlessly integrated by downstream predictors with arbitrary architectures to augment their performances. A series of quantitative and qualitative evaluations on six widely used benchmarks (PEMS03, PEMS04, PEMS07, PEMS08, METR-LA, and PEMS-BAY) are conducted to validate the state-of-the-art performance of STD-MAE. Codes are available at https://github.com/Jimmy-7664/STD-MAE., Comment: Accepted at IJCAI-2024 Main Track

The COVID-19, viral influenza, tuberculosis, and other widespread infectious diseases evidence that pathogenic biological aerosols (PBAs) are a serious threat to public health. Different from traditional inactivation methods, such as ultraviolet (UV) light which are only safe to use when people are not present, and high-efficiency particulate filters (HEPA) which merely filter microbes without killing them, atmospheric pressure nonequilibrium plasma (APNP) has shown its tremendous potential in drastically diminishing the aerosol transmission route of the infectious agents through the abatement of PBAs. The key issues to develop high performance APNP based air purification system are critically reviewed. Systematic studies on the hazards of different PBAs and the spread of PBAs in indoor environments guide the development of APNP sources to control communicable diseases. The key six sampling and seven detection methods on PBAs are introduced to analyze the PBA abatement efficiency by APNP. Seven common APNP sources which can remove viruses and bacteria aerosols efficiently developed during the past 8 years are introduced. For the APNP sources with small plasma volume, the electric field and diffusion driven charging are the dominant mechanisms to charge PBAs, while the common methods of dusty plasma research can be adapted to atmospheric-pressure conditions to describe the charging effects of APNP sources with large plasma volume. Plentiful long- and short-lifetime reactive oxygen and nitrogen species (RONS) generated by APNP effectively contribute to inactivation of bacterial aerosols. Current studies suggest that viral aerosols are mainly inactivated by short-lifetime RONS including 1O2, ONOO- and ONOOH. The study on the dissolution and reaction of gaseous RONS in microdroplets and accurate measurements on the evolution of charged PBAs are envisaged to be the focus of future research. Opportunities for multidisciplinary collaborative research to advance the deve