The field of machine learning is moving towards developing more interpretable models, particularly for tabular data. This is driven by the need for reliability, robustness, and transparency in models used in applications such as finance and healthcare. Recent research has focused on designing models that can provide insights into their decision-making processes, allowing for greater trust and understanding of their outputs. A key direction in this area is the development of models that can learn interpretable features directly from the data, without sacrificing performance. Notable papers in this area include: Preserving Bilinear Weight Spectra with a Signed and Shrunk Quadratic Activation Function, which introduces a new activation function that enables weight-based interpretability. Unveiling the Role of Data Uncertainty in Tabular Deep Learning, which highlights the importance of data uncertainty in explaining the success of recent tabular deep learning methods. An Interpretable Deep Learning Model for General Insurance Pricing, which proposes a model that offers fully transparent and interpretable results for insurance pricing. Towards Interpretable Deep Neural Networks for Tabular Data, which introduces a neural architecture that learns a dictionary of interpretable features for tabular data.