The field of 3D scene reconstruction is witnessing significant advancements with the development of Gaussian Splatting (GS) techniques. Recent research has focused on improving the generalization and expressiveness of GS methods, enabling more accurate and efficient reconstruction of 3D scenes from novel viewpoints. Notably, the integration of neural networks and GS has led to breakthroughs in view synthesis, geometry reconstruction, and dynamic scene rendering. Furthermore, innovations in GS deformation, occlusion culling, and zero-shot image matching have expanded the applicability of GS to various computer vision and graphics tasks. Some noteworthy papers include: Frequency-Adaptive Sharpness Regularization for Improving 3D Gaussian Splatting Generalization, which proposes a novel regularization technique to enhance GS generalization. Neural Texture Splatting is another significant contribution, introducing a global neural field to predict local appearance and geometric fields for each primitive, resulting in improved expressiveness and generalization. Overall, the advancements in GS techniques are poised to revolutionize the field of 3D scene reconstruction, enabling more efficient, accurate, and realistic rendering of complex scenes.