The field of human motion prediction and multimodal sensing is rapidly advancing, with a focus on developing innovative methods that can accurately predict and understand human movements in various environments. Recent research has explored the use of radar-based sensing modalities, which offer robustness and privacy-preserving capabilities, making them suitable for applications such as firefighting and healthcare. Diffusion-based frameworks have also been proposed to improve the accuracy and diversity of human motion prediction models. Additionally, multimodal approaches that integrate audio-visual information are being developed for applications such as dialogue understanding and generation. These advancements have the potential to enable more effective and interactive human-robot collaboration, as well as improve the safety and efficiency of various applications. Noteworthy papers in this area include: mmPred, which introduces a novel diffusion-based framework for radar-based human motion prediction, and SMamDiff, which proposes a spatial mamba-based diffusion model for stochastic human motion prediction. MAViD, a multimodal framework for audio-visual dialogue understanding and generation, is also a notable contribution. Furthermore, GNVC-VD, a generative neural video compression framework, has shown promising results in reducing perceptual flickering artifacts in video compression.