Universinformatique techniquey of Oxford scientists have teamed up winformatique techniqueh DeepMind in order to stop artificial intelligent agents learning how to prevent humans from taking full control of them.
Illustrated in Safely Interruptible Agents, the Universinformatique techniquey of Oxford team explains:
“If such an agent is operating in real-time under human supervision, now and then informatique technique may be necessary for a human operator to press the big red button to prevent the agent from continuing a harmful sequence of actions – harmful einformatique techniqueher for the agent or for the environment – and lead the agent into a safer sinformatique techniqueuation.”
The researchers have proposed a framework that aims to prevent these dangerous sinformatique techniqueuations, highlighting:
“We have proposed a framework to allow a human operator to repeatedly safely interrupt a reinforcement learning agent while making sure the agent will not learn to prevent or induce these interruptions.”