OpenAI last week disclosed new research detailing the training of artificial intelligence (AI) bots to generate their own shared language based on reinforcement learning as they navigate a set environment.
OpenAI scientists tasked the AIs, represented as differently colored circles, with accomplishing specific goals in collaboration by communicating in their own language. The bots produced numerical terms corresponding directly with environmental objects and other bots and actions, which the researchers tagged with English words.
The AIs were taught how to communicate via trial and error, so they could recall successful and unsuccessful strategies the next time they were told to complete a task.
"We think that if we slowly increase the complexity of their environment, and the range of actions the agents themselves are allowed to take, it's possible they'll create an expressive language which contains concepts beyond the basic verbs and nouns that evolved here," the researchers note.
From ReCode
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found