when I was at ISP, I researched interpretable and causal feature representations, focusing on causal inference in human-biosphere systems and developing modules for interpretability and explainability in machine learning models for enhanced understanding.