The field of multimedia compression and semantic communications is experiencing significant advancements, driven by the need for efficient transmission and storage of large-scale multimedia data. Researchers are exploring innovative approaches to improve compression efficiency, reduce latency, and enhance the overall performance of compression systems. One notable direction is the development of hierarchical and progressive context modeling techniques, which enable more efficient exploitation of long-range dependencies and diverse contextual information. Another area of focus is the design of flexible and adaptive compression frameworks that can dynamically adjust to varying device capabilities, power constraints, and channel conditions. Furthermore, the integration of deep learning techniques and transformer architectures is leading to improved performance in image and video compression, as well as semantic communications. Noteworthy papers in this area include: EDPC, which proposes a hierarchically optimized compression framework that achieves comprehensive improvements over state-of-the-art methods. Learned Image Compression with Hierarchical Progressive Context Modeling, which introduces a novel context model that enables more efficient context information acquisition. Hybrid-Domain Synergistic Transformer for Hyperspectral Image Denoising, which proposes a framework that achieves three-dimensional collaborative processing of spatial, frequency, and channel domains. DD-JSCC, which presents a dynamic deep joint source-channel coding approach that enhances the performance of image reconstruction in semantic communications. Rethinking Multi-User Communication in Semantic Domain, which proposes a novel framework that eliminates the need for user-specific JSCC models and enhances privacy. Efficient Sub-pixel Motion Compensation in Learned Video Codecs, which improves learned codec motion compensation by drawing inspiration from conventional codecs. Adjustable Spatio-Spectral Hyperspectral Image Compression Network, which proposes a learning-based model designed for adjustable HSI compression in both spectral and spatial dimensions.