Tag
The paper introduces TTCD, a novel framework for temporal causal discovery from non-stationary time series data using transformer-based feature learning and reconstruction-guided signal distillation.