Scientists make breakthrough in AI navigational ability

LONDON, May 14 (Xinhua) — Scientists have found ways for artificial intelligence (AI) to create navigational ability.

An artificial agent was developed to approximate grid cells, a type of neuron in the brains of many mammal species that allows them to understand their position in space, according to a research report published last week in the journal Nature.

By combining an artificial recurrent network with a larger network architecture, an agent was formed with a deep reinforcement learning ability to navigate itself to goals in challenging virtual reality game environments.

“This agent performed at a super-human level … exhibited the type of flexible navigation normally associated with animals,” said scientists at the London-headquartered Deepmind.

The study tested the theory that grid cells support vector-based navigation, enabling the mammalian brain to calculate the distance and direction to a desired destination.

The results also reflect the philosophy that algorithms used for AI can meaningfully approximate elements of the brain.

“In the future such networks may well provide a new way for scientists to conduct ‘experiments’, suggesting new theories and even complementing some of the work that is currently conducted in animals,” according to Deepmind scientists.




Leave a Reply