Update the Dataset

[VladimirAlexiev]

How up to date is the dataset?

• Eg https://linkedpaperswithcode.com/dataset/gqa lpwc:numberPapers 221 but https://paperswithcode.com/dataset/gqa shows 291 papers
• This query counts papers in the last 10 months. I think LPWC was last updated in 2023-06 (5003 papers in that month); there are only a few in later months:

  PREFIX dct: <http://purl.org/dc/terms/>
  select (count(*) as ?c) ?yearMonth {
  ?x dct:date ?date
  bind(substr(str(?date),1,7) as ?yearMonth)
  } group by ?yearMonth order by desc(?yearMonth) limit 10

  • (Note: the first paper was in 1951)
  • However, https://zenodo.org/records/10056208 is dated Oct 30, 2023

So the delay is 5-5.5 months. But ML research is exploding, so is it possible to implement monthly updates?

I can try to scrounge some resources here at Ontotext to help.

[VladimirAlexiev]

@davidlamprecht wrote "LPWC version v2, which is based on the Papers with code dump from 19.12.2023, by (approx Jan 15) should be loaded into the SPARQL endpoint."

Ok! so that will make less than 1m delay. Thanks!

[VladimirAlexiev]

Running the query today returns 3.7k papers from 2023-12: YAY!