AI for Time Series Causal Discovery in Complex Systems

This research explores the use of AI, specifically deep learning, for causal discovery in time series data in complex systems. Causality is fundamental in various scientific fields, and understanding causal relationships in time series data can lead to important insights and predictions. However, current methods face challenges in high-dimensional datasets and the detection of hidden confounders. The proposal aims to address these issues by exploring new lag selection methods compatible with diverse datasets and investigating efficient deep learning techniques for causal discovery with clear interpretability. The research methodology involves reviewing state-of-the-art approaches, proposing novel models, and comparing them on testing datasets. This research has the potential to contribute to time series analysis, deep learning, causal inference, and statistics. Applications will be in the domain of sustainable development.