It doesn't take much time for most new Minecraft players to work out how to dig up the diamonds that are key to the game, but training artificial intelligence to do it has proved a little tricky.

Recently, Microsoft and other organisations challenged coders to create AI agents that could find the precious gems. Most can crack it in their very first session. But out of more than 660 entrants, not one was up to the task.

The results of the MineRL competition are due to be announced formally on Saturday at the NeurIPS AI conference in Vancouver. The aim had been to see whether the issue could be solved without requiring huge computing power.

Despite the lack of a winner, the organizers were still impressed by some of the participants. Entrants were only allowed to use a single GPU and four days for training. A relatively small Minecraft dataset, with over 60 million frames of recorded human player data, was also made available to entrants to train them

The organizers wanted the coders to make programs that learned by illustration, through a technique known as imitation learning. This involves trying to get AI agents to adopt the most ideal approach by getting them to mimic what humans or other software do to solve a task.

It contrasts with relying solely on reinforcement learning, in which an agent is effectively taught to find the best solution via a process of trial and error, without drawing on past data-collection.

No comments have been submitted to this thread yet.