Learning from lions: inferring the utility of agents from their trajectories

TL;DR

This paper introduces a Gaussian process-based model to infer spatio-temporal utility fields from agent trajectories, revealing influential features and causal factors behind agent decision-making.

Contribution

The paper presents a novel joint Gaussian process framework that models trajectories, utility fields, and influence points, providing deeper insights into agent behavior and decision-making processes.

Findings

Successfully applied to synthetic and lion GPS data

Identified significant influence points affecting trajectories

Enhanced understanding of causal utility functions

Abstract

We build a model using Gaussian processes to infer a spatio-temporal vector field from observed agent trajectories. Significant landmarks or influence points in agent surroundings are jointly derived through vector calculus operations that indicate presence of sources and sinks. We evaluate these influence points by using the Kullback-Leibler divergence between the posterior and prior Laplacian of the inferred spatio-temporal vector field. Through locating significant features that influence trajectories, our model aims to give greater insight into underlying causal utility functions that determine agent decision-making. A key feature of our model is that it infers a joint Gaussian process over the observed trajectories, the time-varying vector field of utility and canonical vector calculus operators. We apply our model to both synthetic data and lion GPS data collected at the Bubye Valley Conservancy in southern Zimbabwe.