S^2-Guidance: Stochastic Self Guidance for Training-Free Enhancement of Diffusion Models

2025-08-18

AI Summary

Classifier-free guidance (CFG) in diffusion models often degrades semantic coherence and generation quality due to its reliance on fixed, deterministic model outputs. Method: We propose S²-Guidance—a training-free guidance strategy that exploits the inherent stochastic subnetwork structure within diffusion models. Specifically, it dynamically applies Gaussian mixture model–driven random block dropping during the forward process to construct lightweight, variable subnetworks, enabling adaptive steering of the denoising trajectory. Contribution/Results: Unlike CFG, S²-Guidance eliminates dependence on static model outputs and requires no auxiliary classifiers, additional training, or external models. Evaluated on text-to-image and text-to-video generation, it significantly improves generation fidelity, prompt adherence, and semantic coherence while demonstrating strong generalization across diverse prompts and architectures.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Classifier-free Guidance (CFG) is a widely used technique in modern diffusion models for enhancing sample quality and prompt adherence. However, through an empirical analysis on Gaussian mixture modeling with a closed-form solution, we observe a discrepancy between the suboptimal results produced by CFG and the ground truth. The model's excessive reliance on these suboptimal predictions often leads to semantic incoherence and low-quality outputs. To address this issue, we first empirically demonstrate that the model's suboptimal predictions can be effectively refined using sub-networks of the model itself. Building on this insight, we propose S^2-Guidance, a novel method that leverages stochastic block-dropping during the forward process to construct stochastic sub-networks, effectively guiding the model away from potential low-quality predictions and toward high-quality outputs. Extensive qualitative and quantitative experiments on text-to-image and text-to-video generation tasks demonstrate that S^2-Guidance delivers superior performance, consistently surpassing CFG and other advanced guidance strategies. Our code will be released.

Problem

• Improving diffusion model output quality
• Reducing semantic incoherence in predictions
• Enhancing prompt adherence without training

Innovation

• Stochastic block-dropping for sub-networks
• Enhances diffusion models without training
• Improves output quality and semantic coherence