Boost-and-Skip: A Simple Guidance-Free Diffusion for Minority Generation

2025-02-10

AI Summary

Generating high-quality samples for minority classes remains challenging due to their sparsity in low-density manifold regions, while existing diffusion-based approaches rely heavily on computationally expensive classifier guidance. Method: This paper proposes an unconditional, lightweight diffusion generative framework. Its core innovations are: (1) variance-enhanced initialization, which strengthens the initial noise representation of minority-class features; and (2) a timestep-skipping mechanism, integrating theory-driven noise covariance rescaling and non-uniform sampling scheduling—achieving improved generation efficiency and fidelity without adding network parameters or external guidance signals. Results: Evaluated across multiple benchmarks, our method surpasses state-of-the-art guided diffusion models in FID, diversity, and other metrics, while reducing computational overhead by over 60%. To our knowledge, it is the first approach to achieve simultaneous breakthroughs in both quality and efficiency for minority-class generation without any guidance.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Web Mining and Content Analysis: Content-based information diffusion
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

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.

Problem

• Generates minority samples using diffusion models
• Eliminates need for computationally expensive guidance
• Enhances minority feature emergence with minimal changes

Innovation

• Guidance-free diffusion model
• Variance-boosted initialization technique
• Timestep skipping strategy