Human-like hippocampus improves self-driving navigation.

Researchers at the Institute of Biophysics of the National Research Council in Palermo, Italy, have developed an AI model that mimics the neural architecture and connections of the hippocampus, allowing a car-like virtual robot to learn and navigate a space more efficiently. The simulated hippocampus can alter its own synaptic connections, meaning it only needs to navigate to a specific destination once before it can remember the path. This is a marked improvement over current autonomous navigation methods that rely on deep learning and have to calculate thousands of possible paths. The system could have applications in industrial settings and helping the visually impaired.