Transflow Learning: Repurposing Flow Models Without Retraining

TL;DR

Transflow Learning is a novel method that repurposes pre-trained generative models to adapt to new data without retraining, using Bayesian inference to warp the latent space for diverse tasks.

Contribution

It introduces a training-free approach to adapt generative models to new data by warping their latent distributions via Bayesian inference.

Findings

Enables style transfer and few-shot classification without retraining

Achieves diverse manipulations by warping latent space

Operates without additional training or fine-tuning

Abstract

It is well known that deep generative models have a rich latent space, and that it is possible to smoothly manipulate their outputs by traversing this latent space. Recently, architectures have emerged that allow for more complex manipulations, such as making an image look as though it were from a different class, or painted in a certain style. These methods typically require large amounts of training in order to learn a single class of manipulations. We present Transflow Learning, a method for transforming a pre-trained generative model so that its outputs more closely resemble data that we provide afterwards. In contrast to previous methods, Transflow Learning does not require any training at all, and instead warps the probability distribution from which we sample latent vectors using Bayesian inference. Transflow Learning can be used to solve a wide variety of tasks, such as neural style transfer and few-shot classification.