SuperAnimal is a revolutionary deep learning model that can automatically identify and track the key point movements of over 45 species of animals without the need for manual annotation. This foundational model is set to fundamentally transform the fields of animal protection, biomedical research, and neuroscience. Through detailed pose analysis, it enables a deeper understanding of an animal's health and motivation, thereby improving animal welfare. It has broad application value in various fie

NASA's Interagency Implementation and Advanced Concepts Team (IMPACT) is jointly developing INDUS, a suite of large language models (LLM) aimed at fields such as Earth science, biological and physical sciences, heliophysics, planetary science, and astrophysics, through signing Space Act agreements with private, non-federal partners. The model is trained using curated scientific literature from diverse data sources.

In recent years, artificial intelligence has been widely applied in the scientific field, but caution is needed to avoid misleading or erroneous results. A new statistical technique called PPI helps researchers safely use machine learning models to validate scientific hypotheses. The PPI technique corrects outputs from large generalized models to fit specific scientific problems, avoiding biases in machine learning. This technology can be applied across various domains, such as protein structure analysis and estimating deforestation in the Amazon rainforest.