Kernel Mean Matching for Content Addressability of GANs
This addresses the need for controllable generation in generative models, offering a novel test-time method for content-addressability, which is incremental as it builds on existing models without fundamental changes.
The paper tackles the problem of adding content-addressability to unconditional implicit models like GANs, enabling users to generate samples similar to a specified set of examples without retraining, and experiments on high-dimensional image datasets show it maintains image quality while achieving consistency with inputs.
We propose a novel procedure which adds "content-addressability" to any given unconditional implicit model e.g., a generative adversarial network (GAN). The procedure allows users to control the generative process by specifying a set (arbitrary size) of desired examples based on which similar samples are generated from the model. The proposed approach, based on kernel mean matching, is applicable to any generative models which transform latent vectors to samples, and does not require retraining of the model. Experiments on various high-dimensional image generation problems (CelebA-HQ, LSUN bedroom, bridge, tower) show that our approach is able to generate images which are consistent with the input set, while retaining the image quality of the original model. To our knowledge, this is the first work that attempts to construct, at test time, a content-addressable generative model from a trained marginal model.