The field of differential privacy is witnessing significant developments, with a growing focus on innovative techniques to protect sensitive information in data science and machine learning applications. Researchers are exploring new methods to provide differential privacy guarantees, such as Gaussian sketching and synthetic data generation, which are showing promising results in improving the trade-off between privacy and utility. Notably, the use of Renyi Differential Privacy (RDP) is leading to tighter bounds and improved performance in various settings. Another area of interest is the development of adaptive algorithms that can handle large-margin linear separable subsets and provide robust privacy guarantees. Theoretical analyses and empirical evaluations are demonstrating the effectiveness of these approaches in maintaining high utility while ensuring strong privacy guarantees. Noteworthy papers include: The Gaussian Mixing Mechanism, which revisits Gaussian sketching through the lens of RDP, providing refined privacy analysis and improved performance in linear regression settings. Adapting to Linear Separable Subsets with Large-Margin in Differentially Private Learning, which obtains an efficient algorithm with empirical zero-one risk bound and improves existing results in the agnostic case.