In this workshop, we aim to bring together emerging research interests within the causal community sharing the goal of learning concise and lower-dimensional representation from data. In recent years, several competing frameworks with similar motivations have been proposed to address these challenges, such as causal feature learning, causal abstraction, causal representation learning, multivariate causal discovery, and cluster DAGs. Despite the growing number of works in these areas, the connections between them remain largely unexplored and under-discussed.
Due to the overlapping motivations of this broad field, we envision this workshop as a joint workshop on causal abstraction, causal representation learning, and related aspects such as multivariate causal discovery. We hope to offer to the community a chance to share intuitions, research ideas, and applications across them. To this end, we will invite submissions of research papers from these areas, giving particular care to works trying to connect intuitions and methodologies. We emphasize that the contributions should contain novel ideas and invite productive discussions, without necessarily being fully developed.