Choice Set Misspecification in Reward Inference

by   Rachel Freedman, et al.

Specifying reward functions for robots that operate in environments without a natural reward signal can be challenging, and incorrectly specified rewards can incentivise degenerate or dangerous behavior. A promising alternative to manually specifying reward functions is to enable robots to infer them from human feedback, like demonstrations or corrections. To interpret this feedback, robots treat as approximately optimal a choice the person makes from a choice set, like the set of possible trajectories they could have demonstrated or possible corrections they could have made. In this work, we introduce the idea that the choice set itself might be difficult to specify, and analyze choice set misspecification: what happens as the robot makes incorrect assumptions about the set of choices from which the human selects their feedback. We propose a classification of different kinds of choice set misspecification, and show that these different classes lead to meaningful differences in the inferred reward and resulting performance. While we would normally expect misspecification to hurt, we find that certain kinds of misspecification are neither helpful nor harmful (in expectation). However, in other situations, misspecification can be extremely harmful, leading the robot to believe the opposite of what it should believe. We hope our results will allow for better prediction and response to the effects of misspecification in real-world reward inference.


Reward-rational (implicit) choice: A unifying formalism for reward learning

It is often difficult to hand-specify what the correct reward function i...

The Effect of Modeling Human Rationality Level on Learning Rewards from Multiple Feedback Types

When inferring reward functions from human behavior (be it demonstration...

LESS is More: Rethinking Probabilistic Models of Human Behavior

Robots need models of human behavior for both inferring human goals and ...

Learning Performance Graphs from Demonstrations via Task-Based Evaluations

In the learning from demonstration (LfD) paradigm, understanding and eva...

I Know What You Meant: Learning Human Objectives by (Under)estimating Their Choice Set

Assistive robots have the potential to help people perform everyday task...

Human irrationality: both bad and good for reward inference

Assuming humans are (approximately) rational enables robots to infer rew...

Avoiding Undesired Choices Using Intelligent Adaptive Systems

We propose a number of heuristics that can be used for identifying when ...