Our paper A Sparsity Principle for Partially Observable Causal Representation Learning was accepted at ICML 2024.