The field of medical imaging and natural language processing is witnessing significant advancements, driven by the development of innovative frameworks and models. Researchers are focusing on creating more robust and efficient systems for image transmission, cancer survival prediction, and language understanding. Notably, the integration of vector quantization and multi-modal learning is improving the accuracy of cancer survival prediction, while novel tokenization frameworks are enhancing the interpretability and generalization of medical imaging models. Furthermore, large language models are being fine-tuned for low-resource languages, demonstrating impressive performance in tasks such as text summarization and question answering. Noteworthy papers include: ResiTok, which proposes a novel framework for ultra-low-rate image transmission that achieves exceptional robustness while maintaining high reconstruction quality. RobSurv, which introduces a robust deep-learning framework that leverages vector quantization for resilient multi-modal feature learning, demonstrating superior performance in cancer survival prediction.