Supervising strong learners by amplifying weak experts
This addresses the challenge of aligning AI with human intent in domains where direct human supervision is infeasible, representing a novel but incremental approach to training strategies.
The paper tackles the problem of training AI systems on complex tasks where objectives are hard to specify directly, by proposing Iterated Amplification, a method that builds training signals from easier subproblems without external rewards, and shows it can efficiently learn complex behaviors in algorithmic environments.
Many real world learning tasks involve complex or hard-to-specify objectives, and using an easier-to-specify proxy can lead to poor performance or misaligned behavior. One solution is to have humans provide a training signal by demonstrating or judging performance, but this approach fails if the task is too complicated for a human to directly evaluate. We propose Iterated Amplification, an alternative training strategy which progressively builds up a training signal for difficult problems by combining solutions to easier subproblems. Iterated Amplification is closely related to Expert Iteration (Anthony et al., 2017; Silver et al., 2017), except that it uses no external reward function. We present results in algorithmic environments, showing that Iterated Amplification can efficiently learn complex behaviors.