Navigation Turing Test (NTT): Learning to Evaluate Human-Like Navigation

TL;DR

This paper introduces an automated Navigation Turing Test (NTT) that predicts human judgments of human-like navigation behavior in 3D environments, enabling scalable evaluation of agent performance.

Contribution

The paper presents a novel automated NTT that learns to predict human assessments, demonstrating high accuracy in distinguishing human from agent navigation behavior.

Findings

Best models achieve high accuracy in classification

Predicting detailed human assessments remains challenging

Automated NTT advances evaluation of human-like navigation

Abstract

A key challenge on the path to developing agents that learn complex human-like behavior is the need to quickly and accurately quantify human-likeness. While human assessments of such behavior can be highly accurate, speed and scalability are limited. We address these limitations through a novel automated Navigation Turing Test (ANTT) that learns to predict human judgments of human-likeness. We demonstrate the effectiveness of our automated NTT on a navigation task in a complex 3D environment. We investigate six classification models to shed light on the types of architectures best suited to this task, and validate them against data collected through a human NTT. Our best models achieve high accuracy when distinguishing true human and agent behavior. At the same time, we show that predicting finer-grained human assessment of agents' progress towards human-like behavior remains unsolved. Our work takes an important step towards agents that more effectively learn complex human-like behavior.