An information-theoretic perspective on intrinsic motivation in reinforcement learning: a survey

  • 2022-09-19 10:47:43
  • Arthur Aubret, Laetitia Matignon, Salima Hassas
  • 0


The reinforcement learning (RL) research area is very active, with animportant number of new contributions; especially considering the emergentfield of deep RL (DRL). However a number of scientific and technical challengesstill need to be resolved, amongst which we can mention the ability to abstractactions or the difficulty to explore the environment in sparse-reward settingswhich can be addressed by intrinsic motivation (IM). We propose to survey theseresearch works through a new taxonomy based on information theory: wecomputationally revisit the notions of surprise, novelty and skill learning.This allows us to identify advantages and disadvantages of methods and exhibitcurrent outlooks of research. Our analysis suggests that novelty and surprisecan assist the building of a hierarchy of transferable skills that furtherabstracts the environment and makes the exploration process more robust.


Quick Read (beta)

loading the full paper ...