Task Force on AI for Time Series and Spatio-Temporal Data

Introduction

Time series and spatio-temporal data are fundamental to understanding dynamic systems across diverse domains such as climate, healthcare, transportation, finance, and urban computing. These data exhibit complex patterns over time and space, often with missing values, non-stationarity, and multi-scale dependencies. Traditional models struggle to capture such complexities, motivating the development of advanced learning methods that can extract meaningful representations, make accurate predictions, and support decision-making.

The growing availability of large-scale temporal and spatial datasets presents new opportunities and challenges for computational intelligence. There is a pressing need for models that are not only accurate and scalable, but also interpretable, robust, and adaptable to diverse data modalities. By advancing learning techniques tailored for these data types, researchers can uncover hidden patterns, support policy interventions, and build intelligent systems that respond to real-world dynamics, ultimately contributing to more resilient, efficient, and equitable outcomes across sectors.

To address these critical needs and drive innovation in this domain, we have established the IEEE CIS Task Force on AI for Time Series and Spatio-Temporal Data—a dedicated initiative focused on advancing research, collaboration, and community building around temporal and spatial learning methods.

Vision: The IEEE CIS Task Force envisions establishing IEEE Computational Intelligence Society (CIS) as a leading hub for advancing learning methods for time series and spatio-temporal data. These data types are central to critical domains such as climate science, healthcare, finance, and smart cities. We aim to foster cutting-edge research in modeling temporal dynamics, spatial dependencies, and other properties like uncertainty.

Mission: Our mission is to promote research and collaboration on computational intelligence techniques tailored for time series and spatio-temporal data. We unify researchers across disciplines, organize community-driven initiatives, and support robust, ethical, and interpretable model development. Through conferences, workshops, benchmarking efforts, and mentoring programs, the task force accelerates progress in both foundational theory and applied solutions. We are committed to building an inclusive community that empowers researchers and practitioners to tackle complex temporal and spatial challenges with real-world relevance and societal value.

Our objectives over a two-year period are listed below:

  • Community Building: Create a vibrant, inclusive community through workshops, webinars, and special sessions at IEEE CIS conferences (e.g., SSCI, WCCI, IJCNN) and other prestigious AI/ML/DM conferences (e.g., ICML, NeurIPS, ICLR, KDD, WWW, AAAI, IJCAI).
  • Knowledge Dissemination: Publish newsletters and organize surveys or special issues in IEEE journals to track emerging trends, benchmarks, and datasets.
  • Interdisciplinary Collaboration: Establish connections with researchers in various domains, such as urban computing, transportation, climate, AIOps, earth science, finance, geography, and public health, and other application domains.
  • Benchmarking & Reproducibility: Promote the development and dissemination of open-source tools, standardized benchmarks, and reproducible pipelines.
  • Next-Generation Talent: Support young researchers through tutorials, PhD mentoring panels, and student competitions.
  • Ethics & Impact: Encourage discussions on fairness, interpretability, and responsible AI in time series and spatio-temporal data.
Major Focuses

This Task Force focuses on the development and application of computational intelligence techniques for modeling, analyzing, and understanding time series and spatio-temporal data. Our key technical focus areas include but are not limited to:

  • Representation learning for time series and spatio-temporal data.
  • Generative methods for temporal and spatial dynamics.
  • Causal inference and intervention modeling for time series and spatio-temporal data.
  • Forecasting and pattern discovery in temporal and spatio-temporal domains.
  • Uncertainty quantification for time series and spatio-temporal applications.
  • Time series foundation models and spatio-temporal foundation models.
  • LLM for processing, analyzing, and modeling time series and spatio-temporal data.
  • Creating datasets and benchmarks for time series and spatio-temporal data.
  • Scalable and efficient learning algorithms for large-scale time series and spatio-temporal data.
  • Interpretable and responsible AI for learning time series and spatio-temporal applications.
  • Cross-domain and multi-modal data fusion methods for time series and spatio-temporal data.
  • Real-world applications of time series and spatio-temporal data in various domains, such as urban computing, transportation, climate, AIOps, earth science, finance, education, geography, and public health.