Umeå University's logo

umu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Expressibility of multiscale physics in deep networks
Umeå University, Faculty of Science and Technology, Department of Physics.
2022 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Motivated by the successes in the field of deep learning, the scientific community has been increasingly interested in neural networks that are able to reason about physics. As neural networks are universal approximators, they could in theory learn representations that are more efficient than traditional methods whenever improvements are theoretically possible. This thesis, done in collaboration with Algoryx, serves both as a review of the current research in this area and as an experimental investigation of a subset of the proposed methods. We focus on how useful these methods are as \textit{learnable simulators} of mechanical systems that are possibly constrained and multiscale. The experimental investigation considers low-dimensional problems with training data generated by either custom numerical integration or by use of the physics engine AGX Dynamics. A good learnable simulator should express some important properties such as being stable, accurate, generalizable, and fast. Importantly, a generalizable simulator must be able to represent reconfigurable environments, requiring a model known as a graph neural network (GNN). The experimental results show that black-box neural networks are limited to approximate physics in the states it has been trained on. The results also suggest that traditional message-passing GNNs have a limited ability to represent more challenging multiscale systems. This is currently the most widely used method to realize GNNs and thus raises concern as there is not much to be gained by investing time into a method with fundamental limitations. 

Place, publisher, year, edition, pages
2022.
National Category
Other Physics Topics Other Computer and Information Science
Identifiers
URN: urn:nbn:se:umu:diva-197523OAI: oai:DiVA.org:umu-197523DiVA, id: diva2:1678353
External cooperation
Algoryx Simulation AB
Subject / course
Examensarbete i teknisk fysik
Educational program
Master of Science Programme in Engineering Physics
Supervisors
Examiners
Available from: 2022-06-29 Created: 2022-06-29 Last updated: 2022-06-29Bibliographically approved

Open Access in DiVA

fulltext(2359 kB)272 downloads
File information
File name FULLTEXT01.pdfFile size 2359 kBChecksum SHA-512
e0b2419a1b0f3a331ec898763144459ab5ac3c08a98009baa91d43e56e577e2a60a7bae8743e6b12a7f1c05e09ba6f473861cd96824d3af5825c462e697c477c
Type fulltextMimetype application/pdf

By organisation
Department of Physics
Other Physics TopicsOther Computer and Information Science

Search outside of DiVA

GoogleGoogle Scholar
Total: 272 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 702 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf