Researchers outline how a robot could be taught to navigate through a maze by electrically stimulating a culture of brain nerve cells connected to the machine. These nerve cells were grown from living cells and acted as the physical reservoir for the computer to construct coherent signals. These findings suggest goal-directed behavior can be generated without any additional learning by sending disturbance signals to an embodied system.

Researchers are challenging a long-held assumption that there is a trade-off between accuracy and fairness when using machine learning to make public policy decisions.

Engineers have built multi-legged robots capable of maneuvering in challenging environments and accomplishing difficult tasks collectively, mimicking their natural-world counterparts.

Researchers have developed a machine learning model to predict never-before-determined material properties from energy loss near-edge structure (ELNES) and X-ray near-edge structure (XANES) spectra. The spectra of 22,155 organic molecules were input into the neural network model and allowed the first correct prediction of both intensive and extensive material properties. It is hoped that the model will allow high-throughput screening to develop materials across numerous applications.

In underwater acoustics, deep learning may improve sonar systems to help detect ships and submarines in distress or in restricted waters. However, noise interference can be a challenge. Researchers now explore an attention-based deep neural network to tackle this problem. They tested two ships, comparing their results with a typical deep neural network, and found the ABNN increases its predictions considerably as it gravitates toward the features closely correlated with the training goals.