Spatio Temporal
Spatio-temporal analysis focuses on understanding and modeling phenomena that evolve over both space and time. Current research emphasizes developing advanced models, such as graph neural networks, transformers, and recurrent neural networks, to capture complex spatio-temporal relationships in diverse data types, including videos, sensor networks, and climate data. These advancements are improving predictions in areas like weather forecasting, traffic flow estimation, and human activity recognition, leading to more accurate and efficient solutions for various applications. The field's significance lies in its ability to extract meaningful insights from complex, dynamic datasets, enabling better decision-making across numerous scientific and practical domains.
Papers
Dynamical system prediction from sparse observations using deep neural networks with Voronoi tessellation and physics constraint
Hanyang Wang, Hao Zhou, Sibo Cheng
StimuVAR: Spatiotemporal Stimuli-aware Video Affective Reasoning with Multimodal Large Language Models
Yuxiang Guo, Faizan Siddiqui, Yang Zhao, Rama Chellappa, Shao-Yuan Lo