Diffusion Based Causal Representation Learning
This addresses the problem of causal reasoning for intelligent systems, offering an incremental improvement over previous VAE-based methods by handling high dimensions and providing infinite-dimensional latent codes.
The paper tackles the challenge of learning causal representations from complex real-world systems by proposing a Diffusion-based Causal Representation Learning (DCRL) algorithm, which performs comparably well in identifying causal structure and variables.
Causal reasoning can be considered a cornerstone of intelligent systems. Having access to an underlying causal graph comes with the promise of cause-effect estimation and the identification of efficient and safe interventions. However, learning causal representations remains a major challenge, due to the complexity of many real-world systems. Previous works on causal representation learning have mostly focused on Variational Auto-Encoders (VAE). These methods only provide representations from a point estimate, and they are unsuitable to handle high dimensions. To overcome these problems, we proposed a new Diffusion-based Causal Representation Learning (DCRL) algorithm. This algorithm uses diffusion-based representations for causal discovery. DCRL offers access to infinite dimensional latent codes, which encode different levels of information in the latent code. In a first proof of principle, we investigate the use of DCRL for causal representation learning. We further demonstrate experimentally that this approach performs comparably well in identifying the causal structure and causal variables.