What paper specifically are you referring to? I couldnt get the paper from the url that was provided, but I am reading the paper from 2017 linked on their github and this project forked ublock, adding additional features, which is pretty interesting actually. Such as detection of visual ads out of the blocked objects, with a series of checks to see if they are "legitimate ads", then simulated clicks through ajax, with blocks on all response content, preserving security from any malware that may have been masquerading as a legit ad.

Granted this is from 2017, but its a pretty interesting idea. https://github.com/dhowe/AdNauseam/wiki/Published-Papers

The code itself was updated three weeks ago, so its clearly still being maintained at some level.

Combine this with containerized user accounts, and seems like a pretty good idea to me.