Unsupervised separation of dynamics from pixels

Silvia Chiappa () and Ulrich Paquet ()
Additional contact information
Silvia Chiappa: DeepMind
Ulrich Paquet: DeepMind

METRON, 2019, vol. 77, issue 2, No 5, 119-135

Abstract: Abstract We present an approach to learn the dynamics of multiple objects from image sequences in an unsupervised way. We introduce a probabilistic model that first generate noisy positions for each object through a separate linear state-space model, and then renders the positions of all objects in the same image through a highly non-linear process. Such a linear representation of the dynamics enables us to propose an inference method that uses exact and efficient inference tools and that can be deployed to query the model in different ways without retraining.

Keywords: Variational auto-encoders; Linear Gaussian state space models; Deep neural networks (search for similar items in EconPapers)
Date: 2019
References: View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
http://link.springer.com/10.1007/s40300-019-00155-4 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:metron:v:77:y:2019:i:2:d:10.1007_s40300-019-00155-4

Ordering information: This journal article can be ordered from
http://www.springer.com/economics/journal/40300

DOI: 10.1007/s40300-019-00155-4

Access Statistics for this article

METRON is currently edited by Marco Alfo'

More articles in METRON from Springer, Sapienza Università di Roma
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().