The long-forgotten technique? Normalizing flows. It feels like it was just last year that they were all the rage. It's insane how fast the field is moving.
this post was submitted on 27 Jun 2025
3 points (66.7% liked)
Hacker News
2254 readers
365 users here now
Posts from the RSS Feed of HackerNews.
The feed sometimes contains ads and posts that have been removed by the mod team at HN.
founded 11 months ago
MODERATORS
It looks like Laurent Dinh (dude who originally came up with normalizing flows) is one of the authors of this work.
There doesn’t seem to be much going on here. What’s the actual advantage of this technique? It seems they’re doing things more or less the same as everyone else.
I mean, the article talks about decoders as if they were some brand new innovation.
Generation ability seems to be about the same as any other model. The advantage of normalizing flows is that operations are invertible. That allows you to not just generate samples, but also calculate the probability of a sample.