We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Statistics > Machine Learning

Title: Model-Free Episodic Control

Abstract: State of the art deep reinforcement learning algorithms take many millions of interactions to attain human-level performance. Humans, on the other hand, can very quickly exploit highly rewarding nuances of an environment upon first discovery. In the brain, such rapid learning is thought to depend on the hippocampus and its capacity for episodic memory. Here we investigate whether a simple model of hippocampal episodic control can learn to solve difficult sequential decision-making tasks. We demonstrate that it not only attains a highly rewarding strategy significantly faster than state-of-the-art deep reinforcement learning algorithms, but also achieves a higher overall reward on some of the more challenging domains.
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Neurons and Cognition (q-bio.NC)
Cite as: arXiv:1606.04460 [stat.ML]
  (or arXiv:1606.04460v1 [stat.ML] for this version)

Submission history

From: Charles Blundell [view email]
[v1] Tue, 14 Jun 2016 17:03:46 GMT (1776kb,D)

Link back to: arXiv, form interface, contact.