Causal Representation Learning via Counterfactual Intervention

Existing causal representation learning methods are based on the causal graph they build. However, due to the omission of bias within the causal graph, they essentially encourage models to learn biased causal effects in latent space. In this paper, we propose a novel causally disentangling framework that aims to learn unbiased causal effects. We first introduce inductive and dataset biases into traditional causal graph for the physical concepts of interest. Then, we eliminate the negative effects from these two biases by counterfactual intervention with reweighted loss function for learning unbiased causal effects. Finally, we employ the causal effects into the VAE to endow the latent representations with causality. In particular, we highlight that removing biases in this paper is regarded as a part of learning process for unbiased causal effects, which is crucial for causal disentanglement performance improvement. Through extensive experiments on real-world and synthetic datasets, we show that our method outperforms different baselines and obtains the state-of-the-art results for achieving causal representation learning.

Authors

• Xiutian Li School of Computer Science, Shanghai Key Laboratory of Intelligent Information Processing, Fudan University, Shanghai 200433
• Siqi Sun Research Institute of Intelligent Complex Systems, Fudan University Shanghai Artificial Intelligence Laboratory
• Rui Feng School of Computer Science, Shanghai Key Laboratory of Intelligent Information Processing, Fudan University Shanghai Collaborative Innovation Center of Intelligent Visual Computing

Keywords

• CV: Interpretability, Explainability, and Transparency
• CV: Representation Learning for Vision