Bayesian Flow Networks in Continual Learning
This work addresses the challenge of continual learning for generative models, which is incremental as it applies an existing method (BFNs) to a new context.
The paper investigates Bayesian Flow Networks (BFNs) for continual learning, tackling the problem of generative modeling on non-stationary data, and empirically verifies their capabilities through experiments.
Bayesian Flow Networks (BFNs) has been recently proposed as one of the most promising direction to universal generative modelling, having ability to learn any of the data type. Their power comes from the expressiveness of neural networks and Bayesian inference which make them suitable in the context of continual learning. We delve into the mechanics behind BFNs and conduct the experiments to empirically verify the generative capabilities on non-stationary data.