The field of biomedical artificial intelligence (AI) is moving towards the development of more advanced and efficient models for named entity recognition (NER) and information extraction. Recent studies have focused on creating large-scale datasets and leveraging transformer-based models and large language models (LLMs) to improve performance. The use of LLMs has shown promising results, often outperforming traditional encoder models, but at a higher computational cost. Additionally, there is a growing interest in developing tools and methods for horizon scanning in healthcare, aiming to improve the efficiency of information retrieval and analysis. Noteworthy papers include:

• A Large-Scale Vision-Language Dataset Derived from Open Scientific Literature to Advance Biomedical Generalist AI, which introduces a large-scale dataset and demonstrates its utility in building embedding models and chat-style models.
• GLiNER-biomed: A Suite of Efficient Models for Open Biomedical Named Entity Recognition, which presents a domain-adapted suite of models for biomedical NER, enabling zero-shot recognition and outperforming state-of-the-art models.