English
 
Help Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
  Learning Structure from the Ground up: Hierarchical Representation Learning by Chunking

Wu, S., Éltetö, N., Dasgupta, I., & Schulz, E. (submitted). Learning Structure from the Ground up: Hierarchical Representation Learning by Chunking.

Item is

Basic

show hide
Genre: Conference Paper

Files

show Files

Locators

show

Creators

show
hide
 Creators:
Wu, S1, 2, Author              
Éltetö, N2, 3, Author              
Dasgupta, I, Author
Schulz, E1, 2, Author              
Affiliations:
1Research Group Computational Principles of Intelligence, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_3189356              
2Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_1497794              
3Department of Computational Neuroscience, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_3017468              

Content

show
hide
Free keywords: -
 Abstract: From learning to play the piano to speaking a new language, reusing and recombining previously acquired representations enables us to master complex skills and easily adapt to new environments. Inspired by the Gestalt principle of grouping by proximity and theories of chunking in cognitive science, we propose a hierarchical chunking model (HCM). HCM learns representations from non-i.i.d sequential data from the ground up by first discovering the minimal atomic sequential units as chunks. As learning progresses, a hierarchy of chunk representation is acquired by chunking previously learned representations into more complex representations guided by sequential dependence. We provide learning guarantees on an idealized version of HCM, and demonstrate that HCM learns meaningful and interpretable representations in visual, temporal, visual-temporal domains and language data. Furthermore, the interpretability of the learned chunks enables flexible transfer between environments that share partial representational structure. Taken together, our results show how cognitive science in general and theories of chunking in particular could inform novel and more interpretable approaches to representation learning.

Details

show
hide
Language(s):
 Dates: 2021-10
 Publication Status: Submitted
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Type: -
 Identifiers: -
 Degree: -

Event

show
hide
Title: Tenth International Conference on Learning Representations (ICLR 2022)
Place of Event: -
Start-/End Date: 2022-04-25 - 2022-04-29

Legal Case

show

Project information

show

Source 1

show
hide
Title: Tenth International Conference on Learning Representations (ICLR 2022)
Source Genre: Proceedings
 Creator(s):
Affiliations:
Publ. Info: -
Pages: - Volume / Issue: - Sequence Number: - Start / End Page: - Identifier: -