Learning Causal Overhypotheses through Exploration in Children and Computational Models

Abstract

Despite recent progress in reinforcement learning (RL), RL algorithms for exploration still remain an active area of research. Existing methods often focus on state-based metrics, which do not con-sider the underlying causal structures of the environment, and while recent research has begun to explore RL environments for causal learning, these environments primarily leverage causal information through causal inference or induction rather than exploration. In contrast, human children - some of the most proficient explorers - have been shown to use causal information to great benefit.In this work, we introduce a novel RL environment designed with a controllable causal structure, which allows us to evaluate exploration strategies used by both agents and children in a unified environment. In addition, through experimentation on both computation models and children, we demonstrate that there are significant differences between information-gain optimal RL exploration in causal environments and the exploration of children in the same environments. We leverage this new insight to lay the groundwork for future research into efficient exploration and disambiguation of causal structures for RL algorithms.

PDF

Resources


Experimental Script

Not Given Hypothesis (Training): Download

Given Hypothesis (Training Video 1): Download

Given Hypothesis (Training Video 2): Download

Reference

If you find this useful in your work please consider citing: 

 
          @inproceedings{kosoy2022clear,
          title = {Learning Causal Overhypotheses through Exploration in Children and Computational Models},
          author = {Kosoy, Eliza and Collins, Jasmine L and Chan, David M, and Hamrick, Jessica B and Ke, Nan Rosemary and Huang, Sandy and Liu, Adrian and Canny, John and Gopnik, Alison},
          booktitle = {Proceedings of Causal Learning and Reasoning (CLeaR)},
          year = {2022}
          }