Can you share the paper on the new method?
EDIT: Perhaps it's this one ? https://arxiv.org/pdf/1711.09025.pdf
That's supervised for one thing. Also this is bot detection, not attempting to detect fake news.
I'm not the first author and I'm not sure what the convention is for pre-prints at WWW.
It's the kind of thing I'd love to talk about because it's so damn amazingly effective, but it isn't really my place to say anything.