Interesting piece and worth a read. Simple answers are always tempting, but in the world of media consumption, they often don’t tell the whole story.

Really important advance in systemically gathered evidence of social media platforms exacerbating & creating disinfo echo chambers

Congrats to Homa & team

These findings go along others in the same line building towards the thought extreme-right groups 'tip the scales' of online debate in much more subtle manners one would think; like a typical complex system, the patterns there may be more elusive than one thinks.

Let me emphasize 3 findings of this new working paper arxiv.org/abs/2011.12843 (with a bunch of co-authors in quoted tweet) (1) News is still a relatively small part of YouTube, but YouTube is huge, news is growing, and right-wing news in particular is growing very fast.

Interested in community detection and link prediction algorithm in complex networks?

Don't miss the talk from Dr. Amir Ghasemian on Monday (11/16) at 8 PM EST.

Details: hamdikavak.com/teaching/2020/…

Amir Ghasemian is a former member of @aaronclauset.bsky.social's lab

How Misinformation Hurts Democracy knowledge.wharton.upenn.edu/article/rothsc… featuring David Rothschild 🌻 #ProjectRatio

A new method for link prediction at PNASNews draws on SFI Prof David Wolpert's #NoFreeLunch Theorems:

'Model stacking performs better than all individual models in most networks considered. And here is the paradox...this looks exactly like a free lunch.'
pnas.org/content/117/41…

Bad time to advertise but I am hiring two positions (Research Data Engineer and Research Project Manager) to help us study media bias and misinformation.

Please apply! Please RT widely! And please, please vote!!

wd1.myworkdaysite.com/en-US/recruiti…

wd1.myworkdaysite.com/en-US/recruiti…

We were invited to write a commentary on the recent paper by Amir Ghasemian, Homa Hosseinmardi, Aram Galstyan, Edo Airoldi and @aaronclauset.bsky.social on model stacking for link prediction in complex networks. This is the result dx.doi.org/10.1073/pnas.2…