David Westergaard

See you on @bluesky !

@ClausEkstrom @HenrikVogt @drjohnbrodersen @bmj_latest Wonderful. Looking forward to read!

Wonderful weekend news. Our paper "How precision medicine and screening with big data could increase overdiagnosis" by @HenrikVogt, Sara Green, @drjohnbrodersen and myself has just been made public by the @bmj_latest bmj.com/content/366/bm…

@AndersPerner @stinuslindgreen Havde samme tanke. Gør det meget svært at sammenligne, specielt landee imellem. Men (måske?) gode nyheder at det bliver opdaget tidligere.

Lige en opløftende nyhed at starte dagen på: Der har været en stor fremgang i overlevelsen ved kræftdiagnoser i Danmark. På tværs af kræfttyper lever vi længere i dag end tidligere. #dkpol #sundpol

@erlichya @MyHeritage @NetanyaMarquis @oweissb Wonderful! What happened with the Github repo mentioned in the first preprint? Will it still be maintained and updated to the version presented in the article?

@ceclindgren @ana_lg_soares @TLaisk @DeborahLawlor2 @don_f_conrad @bmneale @ras_nielsen @jorsmo @ChiayenChen Very impressive work! Pregnancy loss and RPL is an often overlooked condition, and right now its near impossible to stratify these women.

1/n: We formed an international collaboration to dissect the genetic architecture of miscarriage traits: @ana_lg_soares @TLaisk @DeborahLawlor2 @don_f_conrad @bmneale @ras_nielsen @jorsmo @ChiayenChen & many more ⭐️'s biorxiv.org/content/10.110…

@markmccarthyoxf Congratulations Mark!

Some news: after 16+ very happy, productive years in Oxford, I've decided the time has come to move on. It's a wrench to move away from so many fantastic colleagues, and the wonderful students, postdocs, fellows and others who have contributed to our science 1/4

@saralpulit Congratulations!

@eakhram @be_stranger Wonderful review!

Happy to see our review out before the new year! With @be_stranger and Lea K. Davis. Special thanks to Darren Burgess for editorial support.

@f2harrell Thanks for the clarification!

@D_Westergaard The slides I'll be showing are pieces of 5 of the blog articles pointed to with that URL. Don't know if the center is recording this talk.

Excited by the opportunity to give my perspectives on the characteristics, rolls, and accomplishments of #machinelearning vs. statistical models at McGill University in Montreal today: fharrell.com/talk/mlhealth

@wolfgangkhuber Well, in some cases it may have a positive impact on environment, see e.g. static.googleusercontent.com/media/research… - obv theres some costs associated with training initial network, but maybe it can be reused across data centers..

@saralpulit So much to learn, so little time.. Great illustration!

@cshperspectives @cecilejanssens @skathire Cannot agree more! I'm following both for that exact reason

Urge all to follow @skathire @cecilejanssens et al: great (role) models for serious, often actively disagreeing, but always polite sci twitter discourse twitter.com/ethanjweiss/st…

@cecilejanssens Posterior predictive checking :-)

If you fit, say, a regression model on data, when/how do you conclude that the model 'fits' the data?

@erlichya @nevillesanjana @zotero Crap. I was hoping @zotero would be my Mendeley replacement. Latter doesn't work with the Word app from Office365.

@dustinvtran @blei_lab Very interesting work! How does that scale to new cohorts of >100.000 individuals?

Interested in causal models? Check out our work applying it to genomics. #ICLR2018 4:30-6:30p today @ East Meeting (#9). With Dave Blei @blei_lab

@OUPAcademic Link for Supplementary Materials in academic.oup.com/ajcp/article/1… is broken - who to contact? Thanks.

@markmccarthyoxf Very cool projects!

@johnpoverington Good find! It's pretty cool to see a graph CNN in action.