用进化算法在推理时对齐扩散模型，无需梯度或模型内部信息。
Inference-Time Alignment of Diffusion Models via Evolutionary Algorithms
- 将扩散模型视为黑箱，在潜在空间用进化算法搜索最优样本。
- 相同时间内比现有方法高3-35%的ImageReward分数，且节省55%-76%显存。
- 适用于安全约束等目标，适合资源受限场景下的高效对齐应用。
扩散模型是当前最先进的生成模型，但其生成结果常无法满足安全约束或领域特定有效性等应用目标。现有对齐方法通常依赖梯度、模型内部访问或大量计算资源，导致开销高，或不支持某些目标。为此，我们提出一种基于进化算法的推理时对齐框架，将扩散模型视为黑箱，在其潜在空间中搜索以最大化对齐目标。在相同或更短运行时间下，该方法在ImageReward评分上比无梯度和有梯度方法高出3%-35%。在Open Image Preferences数据集上，对四种常见对齐目标均取得竞争性表现。计算效率方面，显存需求减少55%-76%，速度提升72%-80%，显著优于梯度方法。
原文摘要 · Abstract (English)
Diffusion models are state-of-the-art generative models, yet their samples often fail to satisfy application objectives such as safety constraints or domain-specific validity. Existing techniques for alignment require gradients, internal model access, or large computational budgets resulting in high compute demands, or lack of support for certain objectives. In response, we introduce an inference-time alignment framework based on evolutionary algorithms. We treat diffusion models as black boxes and search their latent space to maximize alignment objectives. Given equal or less running time, our method achieves 3-35% higher ImageReward scores than gradient-free and gradient-based methods. On the Open Image Preferences dataset, our method achieves competitive results across four popular alignment objectives. In terms of computational efficiency, we require 55% to 76% less GPU memory and are 72% to 80% faster than gradient-based methods.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。