The current research landscape in the field is characterized by a strong emphasis on developing models that can handle complex, real-world data scenarios, particularly those involving temporal shifts and causal inference. There is a notable trend towards the use of Bayesian approaches and probabilistic models to better understand and predict phenomena affected by a multitude of variables. These models are increasingly being designed to be robust against distribution shifts, leveraging techniques such as in-context learning and structural causal models (SCM) to maintain performance on unseen data. Additionally, advancements in graph-based methods and probabilistic circuits are enabling more efficient and scalable solutions for causal effect estimation and tractable inference. The integration of symmetries in factor graphs and the restructuring of probabilistic circuits are also emerging as key areas of innovation, enhancing the computational efficiency and accuracy of these models. Notably, the development of generative intervention models for causal perturbation modeling is providing new tools for predicting and understanding the effects of external perturbations in complex systems, such as those found in genomics.