Causal blankets: theory and algorithmic framework
File(s)
Accepted version
Author(s)
Rosas De Andraca, Fernando Ernesto
Mediano, Pedro
Biehl, Martin
Chandaria, Shamil
Polani, Daniel
Type
Conference Paper
Abstract
We introduce a novel framework to identify perception-action loops (PALOs) directly from data based on the principles of computational mechanics. Our approach is based on the notion of causal blanket, which captures sensory and active variables as dynamical sufficient statistics—i.e. as the “differences that make a difference.” Furthermore, our theory provides a broadly applicable procedure to construct PALOs that requires neither a steady-state nor Markovian dynamics. Using our theory, we show that every bipartite stochastic process has a causal blanket, but the extent to which this leads to an effective PALO formulation varies depending on the integrated information of the bipartition.
Date Issued
2020-12-18
Date Acceptance
2020-09-15
Citation
Communications in Computer and Information Science, 2020, 1326, pp.187-198
ISBN
978-3-030-64918-0
ISSN
1865-0929
Publisher
Springer
Start Page
187
End Page
198
Journal / Book Title
Communications in Computer and Information Science
Volume
1326
Copyright Statement
© 2020 Springer Nature Switzerland AG.
Source
ECML/PKDD 2020
Publication Status
Published
Start Date
2020-09-14
Finish Date
2020-12-18
Coverage Spatial
Ghent, Belgium