Bayesian Approaches, Distributional Structure

Su discusses the priority of interpretability and modeling semantics, highlighting the learnable mapping between distributional and property spaces. The difficulty in generating the model and the construction of hypotheses are also explored.