The field of diffusion models is rapidly advancing, with a focus on improving image generation quality and developing new methods for inverse problems. Recent research has explored the use of second-order Levenberg-Marquardt-Langevin methods to enhance sampling quality, as well as novel interacting-particle methods for sampling from non-Gaussian distributions. Additionally, there have been significant developments in the use of diffusion models for inverse problems, including the introduction of absorption-based methods and new techniques for posterior sampling. These advances have the potential to improve the accuracy and efficiency of diffusion models in a range of applications, from image generation to protein structure prediction. Notable papers in this area include Unleashing High-Quality Image Generation in Diffusion Sampling Using Second-Order Levenberg-Marquardt-Langevin, which introduces a novel method for improving image generation quality, and Absorb and Converge: Provable Convergence Guarantee for Absorbing Discrete Diffusion Models, which provides a provable convergence guarantee for absorption-based diffusion models. Test-Time Scaling of Diffusion Models via Noise Trajectory Search is also worth mentioning, as it introduces a new method for optimizing noise trajectories in diffusion models, leading to significant improvements in image generation quality.