Literal or Pedagogic Human? Analyzing Human Model Misspecification in Objective Learning

03/09/2019
by   Smitha Milli, et al.
0

It is incredibly easy for a system designer to misspecify the objective for an autonomous system ("robot"), thus motivating the desire to have the robot learn the objective from human behavior instead. Recent work has suggested that people have an interest in the robot performing well, and will thus behave pedagogically, choosing actions that are informative to the robot. In turn, robots benefit from interpreting the behavior by accounting for this pedagogy. In this work, we focus on misspecification: we argue that robots might not know whether people are being pedagogic or literal and that it is important to ask which assumption is safer to make. We cast objective learning into the more general form of a common-payoff game between the robot and human, and prove that in any such game literal interpretation is more robust to misspecification. Experiments with human data support our theoretical results and point to the sensitivity of the pedagogic assumption.

READ FULL TEXT

page 2

page 5

research
09/12/2019

Robots that Take Advantage of Human Trust

Humans often assume that robots are rational. We believe robots take opt...
research
07/06/2021

Physical Interaction as Communication: Learning Robot Objectives Online from Human Corrections

When a robot performs a task next to a human, physical interaction is in...
research
02/03/2020

Quantifying Hypothesis Space Misspecification in Learning from Human-Robot Demonstrations and Physical Corrections

Human input has enabled autonomous systems to improve their capabilities...
research
04/15/2021

Can Artificial Intelligence Make Art?

In two experiments (total N=693) we explored whether people are willing ...
research
10/11/2018

Learning under Misspecified Objective Spaces

Learning robot objective functions from human input has become increasin...
research
06/06/2018

Including Uncertainty when Learning from Human Corrections

It is difficult for humans to efficiently teach robots how to correctly ...
research
05/29/2020

Human-Centric Active Perception for Autonomous Observation

As robot autonomy improves, robots are increasingly being considered in ...

Please sign up or login with your details

Forgot password? Click here to reset