liu.seSearch for publications in DiVA
Operational message
There are currently operational disruptions. Troubleshooting is in progress.
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • 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
A Variational Perspective on Generative Flow Networks
Amsterdam Machine Learning Lab, University of Amsterdam.
Linköping University, Department of Computer and Information Science, The Division of Statistics and Machine Learning. Linköping University, Faculty of Arts and Sciences.ORCID iD: 0000-0003-3749-5820
Amsterdam Machine Learning Lab, University of Amsterdam.
Amsterdam Machine Learning Lab, University of Amsterdam.
2024 (English)In: Transactions on Machine Learning Research, E-ISSN 2835-8856Article in journal (Refereed) Published
Abstract [en]

Generative flow networks (GFNs) are a class of probabilistic models for sequential samplingof composite objects, proportional to a target distribution that is defined in terms of anenergy function or a reward. GFNs are typically trained using a flow matching or trajectorybalance objective, which matches forward and backward transition models over trajectories.In this work we introduce a variational objective for training GFNs, which is a convexcombination of the reverse- and forward KL divergences, and compare it to the trajectorybalance objective when sampling from the forward- and backward model, respectively. Weshow that, in certain settings, variational inference for GFNs is equivalent to minimizing thetrajectory balance objective, in the sense that both methods compute the same score-functiongradient. This insight suggests that in these settings, control variates, which are commonlyused to reduce the variance of score-function gradient estimates, can also be used with thetrajectory balance objective. We evaluate our findings and the performance of the proposedvariational objective numerically by comparing it to the trajectory balance objective on twosynthetic tasks.

Place, publisher, year, edition, pages
2024.
National Category
Probability Theory and Statistics Computer Sciences
Identifiers
URN: urn:nbn:se:liu:diva-204028OAI: oai:DiVA.org:liu-204028DiVA, id: diva2:1863822
Funder
ELLIIT - The Linköping‐Lund Initiative on IT and Mobile CommunicationsWallenberg AI, Autonomous Systems and Software Program (WASP)Swedish Research Council, 2020-04122Available from: 2024-06-01 Created: 2024-06-01 Last updated: 2025-09-24Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

https://openreview.net/forum?id=AZ4GobeSLq

Authority records

Lindsten, Fredrik

Search in DiVA

By author/editor
Lindsten, Fredrik
By organisation
The Division of Statistics and Machine LearningFaculty of Arts and Sciences
In the same journal
Transactions on Machine Learning Research
Probability Theory and StatisticsComputer Sciences

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 209 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • 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