The recent advancements in the field of machine learning and computational modeling have significantly enhanced our understanding of human visual perception and memory. Researchers are increasingly focusing on developing models that can predict and interpret image memorability, leveraging deep learning techniques such as autoencoders and convolutional neural networks. These models are not only improving in their predictive accuracy but also in their ability to identify key visual characteristics that contribute to memorability, such as strong contrasts and distinctive objects. Additionally, there is a growing interest in active learning methods, particularly those that preserve manifold structures, to optimize data annotation processes, thereby reducing bias and improving model performance across various data types. Bayesian approaches are also gaining traction, particularly in concept bottleneck models and hotel booking cancellation prediction, where they offer interpretability and robustness. Furthermore, the integration of human saliency into neural network training is proving to be a powerful method for increasing model interpretability and generalization, with significant improvements observed in both performance and human-model agreement. The field is also witnessing advancements in the interpretation of neural network decision-making processes, with new methods emerging to describe visual concepts using text-image latent spaces. Lastly, the development of specialized Python packages for active learning in regression problems is making these advanced techniques more accessible to researchers across different domains.