The field of multimodal machine translation and understanding is rapidly evolving, with a focus on integrating visual and textual information to improve translation performance and enable more effective communication. Recent research has highlighted the importance of considering the global context of videos and the need for domain adaptation methods to improve performance in out-of-domain scenarios. Additionally, there is a growing interest in developing multilingual multimodal models that can handle diverse languages and cultural contexts. Noteworthy papers in this area include TopicVD, which introduced a topic-based dataset for video-supported multimodal machine translation of documentaries, and Aya Vision, which proposed novel techniques for building multilingual multimodal language models. Other notable works include the introduction of UniEval, a unified evaluation framework for unified multimodal understanding and generation models, and the development of Maya, an open-source multilingual vision language model. These advancements have the potential to drive innovations in multimodal reasoning systems for healthcare scenarios and other applications, ultimately contributing to smarter emergency response systems and more effective education platforms.