<ul data-eligibleForWebStory="true"><li>Structural Causal Models (SCMs) help reason about interventions and support out-of-distribution generalization in scientific discovery.</li><li>Learning SCMs from observed data is challenging, typically necessitating a separate model for each dataset.</li><li>This work introduces amortized inference of SCMs by training a single model on multiple datasets from different SCMs.</li><li>A transformer-based architecture is used for learning dataset embeddings, followed by extending the Fixed-Point Approach (FiP) for SCM inference based on dataset embeddings.</li><li>The proposed method enables the generation of observational and interventional data from new SCMs during inference without parameter updates.</li><li>Empirical results demonstrate the performance of the amortized procedure against baselines, showing competitive results on in and out-of-distribution problems and outperforming them with limited data.</li></ul>

Amortized Inference of Causal Models via Conditional Fixed-Point Iterations

Discover more