Emerging Synergies in Causality and Deep Generative Models: A Survey

Understanding the mechanisms underlying data generation is a fundamental challenge in artificial intelligence. Deep generative models (DGMs) have demonstrated considerable capability in capturing complex data distributions, yet their ability to generalize and provide interpretability remains limited. On the other hand, causality offers a principled framework for explaining data-generating processes by revealing causal-effect relationships. While causality excels in interpretability and the ability to extrapolate, it grapples with intricacies of high-dimensional spaces. Recognizing the synergistic potential, we delve into the integration of causality and DGMs. We provide a comprehensive review of techniques that incorporate causal principles within DGMs, methods for identifying causal relationships through generative modeling, and emerging research frontier of causality in LLMs. We offer insights into methodologies, highlight open challenges, and suggest future directions, positioning our comprehensive review as an essential guide to leverage causal insights to develop more robust and transparent generative models.