Robots could learn from each other
Robots could learn from each other Reuters/Athit Perawongmetha

Soon robots could learn from each other and work together to complete tasks, say researchers.

Artificial intelligence experts at Oxford University have suggested they can teach bots to assess their surroundings and then learn to act on their own.

Supported by the Microsoft Research PhD Scholarship Program, the research was showcased at the AI Summit in London recently.

The researchers hope their AI technology could make it easier for drones to find and rescue people in remote areas ridden with network connectivity issues.

The technology could be used in self-driving cars, as the researchers believe the vehicles could talk to each other about roads and traffic.

Jakob Foerster, one of six researchers, used Microsoft's cloud platform Azure and popular game Starcraft to show (in the video below) how AI bots could learn from their environment and work together.

The video shows two teams of five bots each battling each other and the AI agents losing initially. But as they slowly learn to work together, they are able to win 90% of the games by the time they complete 100 battles.

"What was really interesting was we found that the team got more points when a human player was introduced, rather than just having a group of bots," said Foerster.

"We saw that the human instinct was to attack the big enemies, but the bots were attacking the smaller ones that were causing more damage. The humans then realised what they were doing and started to help the bots. Those were the teams that got the most points," he added.

Drones have previously worked together as a team but it was done with either regular contact with air traffic control or with sensors to monitor the machines' position. The latest research by Oxford shows drones could collaborate without any supervision.