论文标题
要治愈分数匹配的失明
Towards Healing the Blindness of Score Matching
论文作者
论文摘要
基于得分的分歧已被广泛用于机器学习和统计应用。尽管他们的经验成功，但在将它们用于多模式分布时仍观察到了失明问题。在这项工作中，我们讨论了失明问题，并提出了一个新的分歧家庭，可以减轻失明问题。在密度估计的背景下，我们说明了我们提出的分歧，与传统方法相比，报告的性能提高了。
Score-based divergences have been widely used in machine learning and statistics applications. Despite their empirical success, a blindness problem has been observed when using these for multi-modal distributions. In this work, we discuss the blindness problem and propose a new family of divergences that can mitigate the blindness problem. We illustrate our proposed divergence in the context of density estimation and report improved performance compared to traditional approaches.
