基于得分的扩散模型的极大似然训练
Maximum Likelihood Training of Score-Based Diffusion Models
摘要：
基于分数的扩散模型通过反转将数据扩散为噪声的随机过程来合成样本，并且通过最小化分数匹配损失的加权组合来训练。基于分数的扩散模型的对数似然可以通过与连续归一化流的连接来可追踪地计算，但是对数似然不能通过分数匹配损失的加权组合来直接优化。我们表明，对于一个特定的加权方案，客观上界的负对数似然，从而使近似极大似然训练的分数为基础的扩散模型。我们根据经验观察到，极大似然训练一致地提高了基于分数的扩散模型在多个数据集、随机过程和模型架构中的似然性。我们最好的模型在CIFAR-10和ImageNet 32 x32上实现了2.83和3.76 bits/dim的负对数似然，而没有任何数据增强，与这些任务中最先进的自回归模型相当。
英文摘要：
Score-based diffusion models synthesize samples by reversing a stochastic process that diffuses data to noise, and are trained by minimizing a weighted combination of score matching losses. The log-likelihood of score-based diffusion models can be tractably computed through a connection to continuous normalizing flows, but log-likelihood is not directly optimized by the weighted combination of score matching losses. We show that for a specific weighting scheme, the objective upper bounds the negative log-likelihood, thus enabling approximate maximum likelihood training of score-based diffusion models. We empirically observe that maximum likelihood training consistently improves the likelihood of score-based diffusion models across multiple datasets, stochastic processes, and model architectures. Our best models achieve negative log-likelihoods of 2.83 and 3.76 bits/dim on CIFAR-10 and ImageNet 32x32 without any data augmentation, on a par with state-of-the-art autoregressive models on these tasks.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
