Semi-supervised Learning of Visual Causal Macrovariables

• Published in: Proceedings of the International Florida Artificial Intelligence Research Society Conference
• Main Authors: Aruna Jammalamadaka, Lingyi Zhang, Joseph Comer, Sasha Strelnikoff, Ryan Mustari, Tsai-Ching Lu, Rajan Bhattacharyya
• Format: Article
• Language: English
• Published: LibraryPress@UF 2023-05-01
• Subjects: causal representation learning; explainable artificial intelligence; semi-supervised learning
• Online Access: https://journals.flvc.org/FLAIRS/article/view/133229
• ISSN: 2334-0754; 2334-0762

Discovery of causally related concepts is one of the key challenges in extracting knowledge from observational data. Lower-dimensional “causal macrovariables” represent concepts which preserve all relevant causal information in high-dimensional systems. Existing causal macrovariable discovery algorithms are limited by assumptions about known and controllable interventions. We propose a variational autoencoder-inspired architecture with regularization terms for semi-supervised causal macrovariable discovery. These terms impose domain knowledge regarding visual causal concepts to differentiate between correlation and causation. Experiments on both synthetic and real-world datasets with known causal dynamics show that our method can discover more concise and precise causal macrovariables than unsupervised methods.