<ul><li>Long-term causal inference in observational studies is a challenging problem across scientific domains.</li><li>Existing methods address latent confounding in long-term studies using short-term experimental data.</li><li>A new approach suggests a novel assumption that extends previous methods to handle temporal short-term outcomes, enabling identification of long-term causal effects.</li><li>The proposed method is validated through theoretical analysis and experiments, showing its effectiveness in addressing long-term causal inference.</li></ul>

Long-term Causal Inference via Modeling Sequential Latent Confounding

Discover more