What is it about?
Human planning is hierarchical: we first come up with a rough mental sketch of how to achieve a goal and then subsequently refine it on-the-fly. This mental process requires a corresponding hierarchical mental model of the world. In this paper, we provide a computational account of why and how such mental models are built. We find evidence for our proposal across seven behavioral experiments.
Featured Image
Photo by Z on Unsplash
Why is it important?
We find that humans spontaneously build a hierarchical representation of their environment based on the layout of the environment, the rewards they receive, and the tasks they encounter. Furthermore, this process is consistent with principles of optimality. This provides a foundation for studying the neural mechanisms underlying this essential cognitive process and for understanding its disruption in disease states.
Perspectives
I'm excited about this article as it brings together ideas from cognitive psychology, robotics, and animal learning to shed light on a cognitive process that is essential in our everyday thinking and yet remains poorly understood. I hope it inspires future investigations into how our brains represent the world in order to plan and achieve goals so efficiently. I also hope ideas from this work could inspire the development of planning algorithms that match the abilities of humans.
Momchil Tomov
Harvard University
Read the Original
This page is a summary of: Discovery of hierarchical representations for efficient planning, PLoS Computational Biology, April 2020, PLOS,
DOI: 10.1371/journal.pcbi.1007594.
You can read the full text:
Contributors
The following have contributed to this page
