Covered recently in the prestigious journal Nature Medicine, BiomedGPT is a new a new type of artificial intelligence (AI) designed to support a wide range of medical and scientific tasks. This new study, conducted in collaboration with multiple institutions, is described in the article as "the first open-source and lightweight vision–language foundation model, designed as a generalist capable of performing various biomedical tasks."

"This work combines two types of AI into a decision support tool for medical providers," explains Lichao Sun, an assistant professor of computer science and engineering at Lehigh University and a lead author of the study. "One side of the system is trained to understand biomedical images, and one is trained to understand and assess biomedical text. The combination of these allows the model to tackle a wide range of biomedical challenges, using insight gleaned from databases of biomedical imagery and from the analysis and synthesis of scientific and medical research reports."

Researchers from the UCLA Samueli School of Engineering and the University of Rome Tor Vergata in Italy have developed synthetic genes that function like the genes in living cells. The artificial genes can build intracellular structures through a cascading sequence that builds self-assembling structures piece by piece. The discovery offers a path toward using a suite of simple building blocks that can be programmed to make complex biomolecular materials, such as nanoscale tubes from DNA tiles. The same components can also be programmed to break up the design for different materials. 

Research from Case Western Reserve University helps reconstruct an ancient climate and challenges the timing of the Andes Mountains uplift.

Restoring degraded ecosystems has emerged as a global policy priority to address the interlinked concerns of deforestation and land degradation, biodiversity loss and climate change while delivering social benefits, according to the United Nations.

Despite decades of research and meticulous documentation of very small — millimeter-to-centimeter-scale — observations, Utah State University geoscientist Srisharan Shreedharan says reliable, short-term earthquake forecasting remains elusive. To address this challenge, he proposes scaling the wealth of data collected on millimeter-scale faults to the meter scale, using a custom-built, one-meter biaxial deformation “earthquake machine,” along with machine learning techniques. Shreedharan, leading this effort in collaboration with Gregory McLaskey of Cornell University, was awarded about $300,000 in funding, roughly half of the total award, from a three-year National Science Foundation Earth Sciences grant to pursue this approach.