This paper introduces Bayesian Flow Networks (BFNs), a generative model that uses Bayesian inference to modify independent distributions and passes them through a neural network to generate interdependent distributions. BFNs achieve competitive results in image and language modeling tasks.

https://arxiv.org/abs//2308.07037

YouTube: https://www.youtube.com/@ArxivPapers

PODCASTS:

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers