In visual analytics systems, the goal of provenance and interaction analysis is often expressed as the (machine learning) model that a user is constructing, steering, or exploring. In these cases, how the user interacts with the data or the visualization might not be the primary focus and are therefore not directly encoded. Instead, the visual analytics system performs inferencing over the user’s interactions that results in updates to the underlying models. Sometimes known as interactive model steering, in this section we identify publications that: (1) use a sequence of interactions to derive the model, (2) make explicit, quantitative, and recordable representations of these models, or (3) present novel inferencing techniques for analyzing a user’s interactions.We categorize papers in this section into two groups: Machine Learning Models and User Models.