Learning Human Objectives by Evaluating Hypothetical Behavior

By Siddharth Reddy, Anca Dragan, Sergey Levine, Shane Legg, Jan Leike. 2019

83 citations recorded by the source corpus.

We seek to align agent behavior with a user's objectives in a reinforcement learning setting with unknown dynamics, an unknown reward function, and unknown unsafe states. The user knows the rewards and unsafe states, but querying the user is expensive. To address this challenge, we propose an algorithm that safely and interactively learns a model of the user's reward function. We start with a generative model of initial states and a forward dynamics model trained on off-policy data. Our method uses these models to synthesize hypothetical behaviors, asks the user to label the behaviors with rewards, and trains a neural network to predict the rewards. The key idea is to actively synthesize the hypothetical behaviors from scratch by maximizing tractable proxies for the value of information, without interacting with the environment. We call this method reward query synthesis via trajectory optimization (ReQueST). We evaluate ReQueST with simulated users on a state-based 2D navigation task and the image-based Car Racing video game. The results show that ReQueST significantly outperforms prior methods in learning reward models that transfer to new environments with different initial state distributions. Moreover, ReQueST safely trains the reward model to detect unsafe states, and corrects reward hacking before deploying the agent.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Feature Expansive Reward Learning: Rethinking Human Input (2020, background): “For instance, the robot could fit an ensemble of functions from traces online, and query for new traces from states where the ensemble disagrees (Reddy et al., 2019).”
What Would Jiminy Cricket Do? Towards Agents That Behave Morally (2021, background): “Leike et al. (2017); Reddy et al. (2020) investigate reward modeling as a scalable avenue for value alignment.”
Aligning to Social Norms and Values in Interactive Narratives (2022, background): “, 2018) or by modeling rewards that best fit human values (Hadfield-Menell et al., 2016; Reddy et al., 2020).”