9 captures
09 Feb 2022 - 22 Dec 2025
May JUN Jul
05
2023 2024 2025
success
fail

About this capture

COLLECTED BY

Organization: Archive Team

Formed in 2009, the Archive Team (not to be confused with the archive.org Archive-It Team) is a rogue archivist collective dedicated to saving copies of rapidly dying or deleted websites for the sake of history and digital heritage. The group is 100% composed of volunteers and interested parties, and has expanded into a large amount of related projects for saving online and digital history.

History is littered with hundreds of conflicts over the future of a community, group, location or business that were "resolved" when one of the parties stepped ahead and destroyed what was there. With the original point of contention destroyed, the debates would fall to the wayside. Archive Team believes that by duplicated condemned data, the conversation and debate can continue, as well as the richness and insight gained by keeping the materials. Our projects have ranged in size from a single volunteer downloading the data to a small-but-critical site, to over 100 volunteers stepping forward to acquire terabytes of user-created data to save for future generations.

The main site for Archive Team is at archiveteam.org and contains up to the date information on various projects, manifestos, plans and walkthroughs.

This collection contains the output of many Archive Team projects, both ongoing and completed. Thanks to the generous providing of disk space by the Internet Archive, multi-terabyte datasets can be made available, as well as in use by the Wayback Machine, providing a path back to lost websites and work.

Our collection has grown to the point of having sub-collections for the type of data we acquire. If you are seeking to browse the contents of these collections, the Wayback Machine is the best first stop. Otherwise, you are free to dig into the stacks to see what you may find.

The Archive Team Panic Downloads are full pulldowns of currently extant websites, meant to serve as emergency backups for needed sites that are in danger of closing, or which will be missed dearly if suddenly lost due to hard drive crashes or server failures.

Collection: Archive Team: URLs

TIMESTAMPS

The Wayback Machine - http://web.archive.org/web/20240605193817/https://arxiv.org/abs/2202.03813
 
Skip to main content  

Cornell University We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate  



arxiv logo> stat > arXiv:2202.03813  






arXiv logo
Cornell University Logo
 

 

quick links



Login

Help Pages

About
 







Statistics > Machine Learning

 

arXiv:2202.03813 (stat)  



[Submitted on 8 Feb 2022 (v1), last revised 24 Jun 2022 (this version, v3)]

Title:Learning to Predict Graphs with Fused Gromov-Wasserstein Barycenters


Authors:Luc Brogat-Motte, Rémi Flamary, Céline Brouard, Juho Rousu, Florence d'Alché-Buc

View a PDF of the paper titled Learning to Predict Graphs with Fused Gromov-Wasserstein Barycenters, by Luc Brogat-Motte and 4 other authors
 View PDF  
Abstract:This paper introduces a novel and generic framework to solve the flagship task of supervised labeled graph prediction by leveraging Optimal Transport tools. We formulate the problem as regression with the Fused Gromov-Wasserstein (FGW) loss and propose a predictive model relying on a FGW barycenter whose weights depend on inputs. First we introduce a non-parametric estimator based on kernel ridge regression for which theoretical results such as consistency and excess risk bound are proved. Next we propose an interpretable parametric model where the barycenter weights are modeled with a neural network and the graphs on which the FGW barycenter is calculated are additionally learned. Numerical experiments show the strength of the method and its ability to interpolate in the labeled graph space on simulated data and on a difficult metabolic identification problem where it can reach very good performance with very little engineering.  

Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:2202.03813 [stat.ML]
  (or arXiv:2202.03813v3 [stat.ML] for this version)
  https://doi.org/10.48550/arXiv.2202.03813




Submission history

 From: Luc Brogat-Motte [view email] 
[v1]  Tue, 8 Feb 2022 12:15:39 UTC (1,703 KB)
 [v2]  Wed, 16 Feb 2022 11:49:22 UTC (1,507 KB)
 [v3]  Fri, 24 Jun 2022 14:37:01 UTC (1,706 KB)

 

Full-text links:  

Access Paper:



View a PDF of the paper titled Learning to Predict Graphs with Fused Gromov-Wasserstein Barycenters, by Luc Brogat-Motte and 4 other authors
View PDF
TeX Source
Other Formats

view license
 

Current browse context: 
stat.ML

< prev     |    next >  

new  |  recent  | 2022-02  
Change to browse by:  
cs
cs.LG
 stat
 



References & Citations



NASA ADS
Google Scholar

Semantic Scholar
 



export BibTeX citation  



BibTeX formatted citation

 ×  



Data provided by: 



Bookmark

 BibSonomy logoReddit logo

 

Bibliographic and Citation Tools






Bibliographic Explorer (What is the Explorer?)  





Litmaps (What is Litmaps?)  





scite Smart Citations (What are Smart Citations?)  





 

Code, Data and Media Associated with this Article






CatalyzeX Code Finder for Papers (What is CatalyzeX?)  





DagsHub (What is DagsHub?)  





Gotit.pub (What is GotitPub?)  





Papers with Code (What is Papers with Code?)  





ScienceCast (What is ScienceCast?)  



 

Demos






Replicate (What is Replicate?)  





Hugging Face Spaces (What is Spaces?)  





TXYZ.AI (What is TXYZ.AI?)  






 

Recommenders and Search Tools






Influence Flower (What are Influence Flowers?)  





Connected Papers (What is Connected Papers?)  





CORE Recommender (What is CORE?)  


  • Venue
  • Institution
  • Topic







  • arXivLabs: experimental projects with community collaborators


    arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

    Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

    Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
     








    Which authors of this paper are endorsers? |  Disable MathJax (What is MathJax?)  








    About

    Help
     




    contact arXivClick here to contact arXiv  Contact  

    subscribe to arXiv mailingsClick here to subscribe  Subscribe  









    Copyright

    Privacy Policy
     




    Web Accessibility Assistance


    arXiv Operational Status 
    Get status notifications via  emailorslack