Optimizing Trust and Safety Regions for Text-to-Image Generation in High-Dimensional Manifold Spaces
Abstract
Diffusion-based text-to-image (T2I) models such as Stable Diffusion (SD) and DALL$\cdot$· E 2 enable versatile image generation but raise significant safety concerns due to their ability to produce harmful or not-safe-for-work (NSFW) content (e.g., nudity). Existing safety strategies, including prompt filtering and machine unlearning, remain limited, as they are vulnerable to biased data, model openness, and adversarial prompt attacks. Achieving safe alignment during reinforcement learning (RL) fine-tuning is thus essential, yet faces two significant challenges: alignment fragility, where models easily lose control after optimization, and the safety-quality paradox, where improving safety often degrades visual quality. To address these issues, we propose S-TRPO, a Safety-constrained Trust-Region Policy Optimization framework that enables safe and reliable alignment of diffusion models (DMs) within the manifold policy space. S-TRPO introduces a dynamic safety-control mechanism that combines danger-region perception with trust-region constraints to maintain both safety and generation fidelity. Specifically, a KL-based safety region and a static risk model jointly evaluate harmful prompt risk and restrict unsafe deviations in policy updates. Furthermore, a Lagrangian dual-control scheme balances safety constraints with image-quality optimization. Extensive experiments on real-world adversarial benchmarks demonstrate that, under white-box UnlearnDiffAtk evaluation, S-TRPO with…
Read the paper
Related studies
- Learning from First-Generation Qualitative Approaches in the IS Discipline: An Evolutionary View and Some Implications for Authors and Evaluators (PART 1/2)
- Commentary on the clinical and preclinical dosage limits of interstitially administered magnetic fluids for therapeutic hyperthermia based on current practice and efficacy models
- Prosthesis: New Technological Opportunities and Innovative Biomedical Devices
- The history and future of neural modeling for cochlear implants
- Ryanodine receptors: waking up from refractoriness
- Novel Transgenic Rabbit Model Sheds Light on the Puzzling Role of Matrix Metalloproteinase-12 in Atherosclerosis
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.