The field of face generation and animation is moving towards greater controllability and personalization. Researchers are exploring new methods to achieve fine-grained control over facial features, such as identity, attributes, and age, without requiring extensive training data or additional modules. One notable direction is the use of pre-trained expert models to guide the generation process, allowing for more accurate and realistic results. Another area of focus is the development of plug-and-play components that can be easily integrated into existing systems, enabling more flexible and efficient generation of high-quality talking head videos and 3D speech-driven facial animations. These advancements have the potential to revolutionize applications such as digital avatars, online education, and customer service. Noteworthy papers include ExpertGen, which leverages pre-trained expert models for controllable face generation, and FaceEditTalker, which enables interactive talking head generation with facial attribute editing. Additionally, RESOUND and Wav2Sem propose novel approaches to speech reconstruction and audio semantic decoupling, respectively, demonstrating significant improvements in accuracy and naturalness.