Facial self-touching — particularly around the nose, chin and cheeks — strongly correlates with elevated stress during cognitive tasks, new University of Houston research showed.

This study used the concept of reinforcement learning to explain the navigation of chemotactic cells toward sparsely distributed targets, showing how decentralized information processing through environmental interaction can lead to highly intelligent behavior. Simulations showed that groups of simple agents could navigate mazes more robustly than a more intelligent single agent. This demonstrates that decentralized teams of simple agents can efficiently process information as a group, with potential applications in medicine, artificial intelligence, and robotics.

A Harvard team has demonstrated that robots can be designed to react to their environment and perform tasks by programming intelligence into their structure. 

AnalySwift LLC, a composite simulation software provider, has launched a new artificial intelligence platform to help users more quickly develop composite materials and structures without requiring in-depth technical jargon or knowledge. CEO Allan Wood said one application of the free platform is for the air mobility industry to design composite rotor blades and propellers. AnalySwift partnered with Wenbin Yu, the Milton Clauser Professor of Aeronautics and Astronautics at Purdue University, to create CompositesAI.

The Princeton Plasma Physics Laboratory will provide critical measurement technology for the world’s largest fusion machine, JT-60SA. It is among the first contributions to the fusion system from an institution outside of Japan and Europe.

A research team has reviewed how machine learning (ML) is revolutionizing fermentation design and process optimization by providing powerful simulation and prediction tools.