Saved in:
Bibliographic Details
Main Authors: McKee, Kevin, Alt, Eric, Grebenisan, Andrew, van Gelderen, Mick, Miguel, Gary
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.02831
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916975335178240
author McKee, Kevin
Alt, Eric
Grebenisan, Andrew
van Gelderen, Mick
Miguel, Gary
author_facet McKee, Kevin
Alt, Eric
Grebenisan, Andrew
van Gelderen, Mick
Miguel, Gary
contents Exploration algorithms for reinforcement learning typically replace or augment the reward function with an additional ``intrinsic'' reward that trains the agent to seek previously unseen states of the environment. Here, we consider an exploration algorithm that exploits meta-learning, or learning to learn, such that the agent learns to maximize its exploration progress within a single episode, even between epochs of training. The agent learns a policy that aims to minimize the probability density of new observations with respect to all of its memories. In addition, it receives as feedback evaluations of the current observation density and retains that feedback in a recurrent network. By remembering trajectories of density, the agent learns to navigate a complex and growing landscape of familiarity in real-time, allowing it to maximize its exploration progress even in completely novel states of the environment for which its policy has not been trained.
format Preprint
id arxiv_https___arxiv_org_abs_2503_02831
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Meta-Learning to Explore via Memory Density Feedback
McKee, Kevin
Alt, Eric
Grebenisan, Andrew
van Gelderen, Mick
Miguel, Gary
Machine Learning
Exploration algorithms for reinforcement learning typically replace or augment the reward function with an additional ``intrinsic'' reward that trains the agent to seek previously unseen states of the environment. Here, we consider an exploration algorithm that exploits meta-learning, or learning to learn, such that the agent learns to maximize its exploration progress within a single episode, even between epochs of training. The agent learns a policy that aims to minimize the probability density of new observations with respect to all of its memories. In addition, it receives as feedback evaluations of the current observation density and retains that feedback in a recurrent network. By remembering trajectories of density, the agent learns to navigate a complex and growing landscape of familiarity in real-time, allowing it to maximize its exploration progress even in completely novel states of the environment for which its policy has not been trained.
title Meta-Learning to Explore via Memory Density Feedback
topic Machine Learning
url https://arxiv.org/abs/2503.02831