Boost-and-Skip: A Simple Guidance-Free Diffusion for Minority Generation
This work addresses the problem of minority sample generation for researchers and practitioners in generative AI applications, such as data augmentation and creative content generation, providing an efficient alternative to existing guidance-based methods.
The authors tackled the problem of generating minority samples in data manifolds, achieving results that rival state-of-the-art guidance-based approaches while requiring significantly fewer computations. Their Boost-and-Skip method demonstrates enhanced capability in generating minority samples.
Minority samples are underrepresented instances located in low-density regions of a data manifold, and are valuable in many generative AI applications, such as data augmentation, creative content generation, etc. Unfortunately, existing diffusion-based minority generators often rely on computationally expensive guidance dedicated for minority generation. To address this, here we present a simple yet powerful guidance-free approach called Boost-and-Skip for generating minority samples using diffusion models. The key advantage of our framework requires only two minimal changes to standard generative processes: (i) variance-boosted initialization and (ii) timestep skipping. We highlight that these seemingly-trivial modifications are supported by solid theoretical and empirical evidence, thereby effectively promoting emergence of underrepresented minority features. Our comprehensive experiments demonstrate that Boost-and-Skip greatly enhances the capability of generating minority samples, even rivaling guidance-based state-of-the-art approaches while requiring significantly fewer computations. Code is available at https://github.com/soobin-um/BnS.